ADVANCE Live Reporter user manual
Contents
1. 8zz Ln 8JgIstA Jeny zz Ln 8JqIstA Jeny i i jemeds s z OLE ogl 12303 Jeny l l Jepads m OLE osi 12303 jeny ErL 99 Duoud E a zz Loz SIGISIA paa2tpaud F 99 Duoud E B zz LOZ IqISIA P J3IP 1d SIZ ZOL Piepuej s m ut 6SE 1230 payipaid GZ ZOL psepuejs m GI 6SE 1210 paqipalg jee papayas ysaqe pejeles jee paypajas jee papayas o Wd Wd9 Wat Wdr WdE Wdt Wd Wd Wv went We HM HIM Wd Wd Wd9 Was Wd Wat Wdz Wdl Wade Weit wvol Wee HIM ro 0 1 ro t 31 D 9 97 9 6 6 1 6 HI UT a E Gt 1 Sr s SS si 811 ee EET D t eel ul u u PEPE wz L L v Lt lt z 0t D 0t Ka 9t 9 6t 6t 6t y d L ae 3od q I punoqui 3od q punoqui 20 A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble 3 5 Administrator functions Note regard these functions in their present form as preliminary User comments from field tests are still welcome for these functionalities and further changes may be carried out if they are requested or justified by feedback Currently the administrator acts as a hub user with special privileges i e has access to hub views and can change user and configuration settings In its current version the ALR offers the following administrator specific functions via the a drop down menu opened with the Setting2. pe1ea1 peuue g Page deo Aeq 0 Bil puepueis 3ndui eu491X3 suonejnoje2 Mou qnu iy deo Aeq oi Si jelseds oud mou peo Keg ndu euj93x3 deo Aeq oi ail piepueis mou peo Aeg 3ndu euje1x3 50 A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble Appendix B Consolidation of warehouse layouts and ca pacities B 1 Motivation Generally warehouse pairs are laid out to partition the existing bays and A and B warehouses are meant to be identical copies of each other Exceptions however may occur to the congruence of A and B e The order and location of bays may not be the same in A and B e Some smaller bays may be exclusive to either A or B e Capacities of the same bays may differ between A and B It is fairly easy to handle different capacities individually but the warehouse topology i e the order of bays and whether they are in the left or right row needs to be consolidated into a merged model to allow a side by side view of the same bays in A and B in the eve 2 warehouse view The warehouse layout editor see Sec 3 5 2 p 21 allows differences between A and B and a consolidated layout as used in the eve 2 warehouse screens is generated automatically Next we take a closer look at how this unified layout is generated B 2 Merging warehouse layouts B
3. Common screen elements Level 1 global warehouse overview Level 2 bay overview Level 3 bay details Deliverable A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable Appendix A Bars and colours in warehouse screens 36 A1 Motivati n sesos a agos iaa ae df aa ia a a E kae a E paa 36 A 2 Saturation in bar lengths 4 oa aaa oy GaSe ae REY 36 A 3 Generating colours amp 244 4448 4 doe 4 44444444444 4 644445 37 A 3 1 Colour transitions uox Aussen en RIS 39 A 3 2 Adjusting input sensitivity 1 2 64 c ok os 40 AA Inputs and specific parameters oo a a a a 42 PAN At hub now DAES 52k 9 ov RR o9 ooh 44 805 44845 42 A 4 2 Coming up bars oae Be Po eR he EUR 43 A 4 3 The overall state dot 2 2 a ee Ar mes Rx ee Does eRe 44 Ab Putting it all together uuu sure oh Ose obo de od OR Rehd 45 Appendix B Consolidation of warehouse layouts and capacities 51 E M tivati h su nde eee Se xA e BS Bed M Q eee Soe Xu oe WQ 51 B 2 Merging warehouselayouts e 51 B 3 Handling capacities 2 6 26 Ho So RGE S a 54 A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble Executive summary This supplement to D5 3 covers the functionalities of the ADVANCE Live Reporter as imple mented to complement the OS release of the ADVANCE framework Specifically the following areas are ex
4. 0 1 Same for both service types Table A3 Parameters for coming up bar length and background colour functions A note on inputs for the weighted sums these are as everything else in the warehouse screens ignoring predicted values in the current implementation of the Live Reporter Once the prediction method proves to be reliable it is planned that all warehouse screens will ignore scanned and created states and use the predicted value instead of these Once this change is carried out the weighted sums in the coming up bars will be calculated using the declared values with the coefficient Wm and the predicted values with the coefficient wy A 4 3 The overall state dot The overall state dot next to the bay number in the level 2 warehouse screens summarises the state of affairs regarding the given bay in general Here we take the inputs determining the at hub now and coming up background colours both being maxima of relative loads a weighted sum in the coming up case and form their weighted sum using the coefficients zu and we respectively This weighted sum is the input for the usual colour function Weighting factors and parameters for the overall state dot are listed in Tab A 4 Parameter Current value Underfull sensitivity p 0 85 Overfull sensitivity q 1 2 Weight for at hub now maximum w 0 75 Weight for coming up maximum w 0 25 Table A4 Parameters for the colour of t
5. 1 Depot 1 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable The unit of measurement selector specifies what units the given traffic is displayed in In the current implementation these can be price number of separate shipping units and occupied floorspace P E AD VANCE Live Reporter x Next days During day 1 Depot 1 DESS Count Floor Figure 3 Header and common controls for hub mode screens 3 2 The Summary screen The Summary screen gives roughly speaking an overview of all shipments passing through the hub in the current shift The amounts are shown for the entire hub or for a selected depot both inbound and outbound in the latter case in the unit of measurement selected by the user The screen shows the total amounts for all service types as well as separate numbers for Standard Priority and Special For the correct interpretation of the diagrams it is vital to understand how shipment states are defined In a simplified approach each shipment passes through following states see Fig 4 for colours 1 Created the shipment is entered into the system receives a unique identifier and becomes visible to the system 2 Scanned the shipment is scanned onto a vehicle that will at some point depart form the depot of origin 3 Declared once a shipment is known to be declared the depot of
6. declared amount displayed In addition to numbers known to the system the ALR adds a during day prediction to the diagram The predicted number shows how much more can be expected by the end of the shift in addition to the shipments already declared On occasion this can be less than the sum of created and scanned shipments and this may be very well in line with the reality of a given business case not all created or scanned shipments would actually get declared at all or by the end of the day Left hub At hub Declared Scanned Created i Predicted Figure 4 Colour codes of different shipment states as shown in legend areas of the Summary screen Having clarified the meaning of shipment states let us return to the diagrams shown in the Summary screen Fig 5 In the case of the hub overall we see four compound bars standing for all services Standard Priority and Special respectively The coloured areas of the bars correspond to shipment states as defined above with successive states moving from right to left In addition to the graphical display the exact amounts are also shown in numbers in the legend located to the right of each compound bar The prediction how much more is to be expected than those already manifested is shown as a dashed outline and may overlay or go beyond the scanned and created bars Two additional features support the user in relating bar lengths to numbers rulers and pop up number boxes For
7. uiue jeg u13ue jeg puepueas u13ue jeg u13ue jeg seniAnisues DUN ino o2 dn 3uiuo 104 s Aanisuas q3us jeg dn Sue 104 S44319M snieis 19 ed peqeas peuueog p 4e 2 q deo Aeq o ail epadet ou 3ndui euje3x3 peqeas pauuess p 4e q de Aeq o je4 puepuers 3ndui euj 1xg 48 Deliverable ueap aow jo p V DBS 99s JOP 2424S e49 0 BY 104 suonqe no eO 6 V eni 1nojo2 punojSy eg JOP ne3s Weieng 2 e49AQ 4 e4240 suone naje5 2424S amp J9AQ bd 81815 amp 49AQ 104 seniAnisues DUN INO O7 MOU qny 3V pue dn Suiwos JO S1u3I9M sn1e1S Advanced Predictive Analysis Based Decision Support Engine for Logistics AD VANCE 49 Deliverable Advanced Predictive Analysis Based Decision Support Engine for Logistics AD VANCE U9945S BSNOYaJeEM Z aAa y Jo s1u uu 01 UOnEIa 1194 pue suolle no e2 sn1e3s Aeq UE jo uone es OT V eJn2i4 suone na e dn Zulwod pe1ea1 peuue g paie peq deo Aeq oi o el2 ds ouq ndu euj 1xd Suonejn je 1e3s amp 49AQ
8. 10096 Figure A4 Mapping of relative load to colour control variable t for loads over 100 41 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable The equations for the curves in Fig A 4 are b 7 EE 8 where q is the sensitivity parameter should always be greater than 1 and is set to 1 2 in the current Live Reporter implementation z is the relative load and is the output passed on for generating the background colour Decreasing q results in a faster response with the display turning black at smaller overloads while increasing q puts off the response to higher overloads Load 10096 Name of variable Formula used Red colour component r 3 Green colour component g 4 Colour control variable t 5 6 Load gt 100 Name of variable Formula used Red colour component r 3 Green colour component g constant 0 Colour control variable t 7 8 Table A1 Formulae applied in underfull top and overfull bottom cases AA Inputs and specific parameters A 4 1 At hub now bars The inputs for the at hub now bars are relative floorspace occupation data based on which shipments are known to be in the given bay in the given warehouse separately for Standard and Priority 4 NextGen Floorspace occupation is at the moment more an estimation than exact information as the data model currently in use d
9. a password to be entered see Fig 1 A given user is assigned to one specific user group this also determines 7 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics De I Ivera b le access rights as well as the set of screens the user will see ADVANCE Live Reporter Username admin Password eeccccccce iD Sign in Figure 1 ALR login screen with input fields filled out by user Entering an invalid login and password pair clears the input fields and brings up a warning message below the fields ADVANCE Live Reporter User name Password emu Sign in Unable to log you in Invalid user name or password Figure 2 ALR login screen upon entering invalid input A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble 3 Screens in hub mode The hub mode screens are meant for users overlooking hub activities on hub level i e not going into details of individual warehouse bays etc Hub mode screens are designed for a desktop computer with e adequate screen size and resolution and e a mouse or similar pointing device that allows hovering the cursor over screen elements In general screens have been designed for minimal user interaction clicking scrolling etc where gaining a fast overview is of importance Also selected display options are preserved while the user swit
10. basis for sorting bays with the option The bay status bars are accompanied by a common header in accordance with the symmetrical arrangement of left and right bay rows the structure of these also forms a pair of mirror images Here the header of the left row is explained e Above the rightmost coloured field Coming up signalises that this field is for shipments that have not been tipped yet e At hub now stands above the field for current bay load e The warehouse name is shown above the bay number and the round dot for the general bay state 32 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics De I Ivera b le AN AQ Live Reporter Warehouse 1 A A w Warehouse 2 L R B At hub now Coming up Coming up At hub now B Oe e e E e eS www Ween i 2 o M E w nn e e HEN E u nwaa lt aH e i E sa O0 s w H OrI 7 www ra w 170 24 w w NT 0 o n ws Oo r m 0 SSS C CU 560 Figure 21 Example of a level 2 screen showing all bays for a selected warehouse 33 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli vera ble In the Level 2 view the level changer green triangle in the top left corner points upward tapping on this control will bring the user back to Level 1 Level 3 view of a given bay can be opened by tapping o
11. bay capacities can very well be adjusted to actual values also with regard to economy and premium usage of floorspace e Simplified mathematics behind colours and bar lengths indicating severeness of upcoming traffic is due to the fact that detailed information on upcoming resources e g trunking schedule vehicle capacities vehicles on site is not yet sufficiently reliable for building a consistent and more sophisticated model for judging a situation The implementation of the screens is however open for future additions once more reliable data become available Shed screens represent in their current form an initial step in supplying warehouse personnel with better information in a readily available form i e extending the current possibilities of on site communication and much practical experience will have to be gathered before moving on to further functionalities such as managing vehicles 23 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli vera ble A note on units everything in the warehouse screens is measured by footprint and not billing units or lifts since floorspace occupation is the key concern when warehouse level decisions are met This includes the capacity of bays which is also measured in shipment floorspace even if floorspace occupation is in some cases expressed relative to bay capacity 4 1 Common screen elements All warehouse screens share a common hea
12. cap x bay cap 7 Warehouse 1 Depot 7 Figure 23 Example of a level 3 screen note the scroll bar to the right indicating that all bays in the same row can be browsed by scrolling 35 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable Appendix A Bars and colours in warehouse screens A 1 Motivation Shed overview screens have to provide easy to recognise status information within very limited display real estate Although a plain numerical display would be a very space efficient solution this is certainly not favoured in the given application environment where fast orientation and recognition of a critical situation is required at a single glance Therefore a system of bar lengths and background colours was designed for the warehouse overview screens where bar lengths stand for floorspace quantities per service type while background colour signalises the overall situation Certainly a graphical display seems a straightforward choice however we have to keep in mind that the available real estate for a given bar is limited and the colour scale is finite while there is no theoretical limit to the amount of shipments or floorspace over utilisation in bays This calls for bar lengths and colour transitions that saturate and thus prevent off the scale display states Practically this would mean that anything up to a nominal 10096 mark would be shown in better deta
13. choice for bays with outstanding turnover Therefore the current implementation leaves the door open for further improvement upon a more elaborate analysis of tipping and loading dynamics e The rectangular field to the left of the coming up bars shows the current floorspace occupation relative to bay capacity The white bars represent Standard top and Prior ity Special bottom floorspace occupation with the length of the bars going gradually into saturation when overload occurs With the current sensitivity settings 10096 bay oc cupation corresponds to 85 of maximal available bar length As with the coming up bars the colour of the background rectangle shows the maximum of relative occupation in the Standard and Priority Special slots either an Standard or a Priority Special space overfilled at 120 is more severe than both slots being say 75 full Green corresponds to an empty bay red stands for 10096 occupation while overloaded bays gradually turn 31 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable black e Next to the at hub now field is the bay number and a round dot representing the general state of the bay with its colour The latter is essentially a weighted sum of current and upcoming values i e the worst load values determining the background colour of the fields with the current load receiving the strongest weight It is this measure that forms the
14. matched at the 100 mark red but with sensitivity characteristics that can be set separately In our implementation we took a two step approach to doing this e First we set up both colour transitions with mathematical functions that accept an inter mediate variable t which is 0 0 at one end of the transition and 1 0 at the other end What t actually means depends on whether we are below or above 100 load e Next we set up two functions that translate relative load into t This is the point where we can introduce sensitivity parameters Also here we will have two separate functions with independent sensitivities Which one of these is actually used to calculate will depend on how the input relates to the 10096 threshold As it can be seen a switching between functions is required at the 100 limit both for generating the colours and for calculating the colour control variable t Next we present the colour functions and the control variable calculation separately finally Tab A 1 at the end of this section p 42 will summarise which functions are to be used in different cases 37 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics De I Ivera b le Value of sensitivity parameter s 0 95 0 85 0 75 Output bar length in of maximal bar length 0 0 5 1 1 5 2 2 5 3 Input footprint relative to nominal 100 Input Displayed bar length with s 0 75 0 85 0 95 0 2
15. numbers Note that scaling remains the same the highlighter bar remains at the same place and service type visibility options are preserved facilitating the comparison of inbound and outbound graphs Colour codes for inbound graphs are consistent with those in the Next days screen AQ E ADVANCE Live Reporter bed Summary Next days 1 Depot 1 DESS Count Floor EC JESS Inbound 1 Depot 1 x40 Too 42 2 39 39 h U 36 36 33 30 27 24 t21 t18 15 Ai 9 3 30 27 24 n 18 15 2 9 6 3 6 3 di TPM 8PM x10 Selected Latest ol xt0 9AM 10AM MAM 12PM 1PM 2PM Selected Latest 6PM Predicted total 375 372 B Standard 157 215 Predicted visible 375 372 Priority 85 143 Actual total 255 370 E Special 13 13 Actual visible 255 370 Lei A E ADVANCE Live Reporter Summary Next days 1 Depot 1 DESS count Foor ibd Outbound 1 Depot 1 x10 y x0 42 4 39 39 36 36 33 33 30 30 x x 241 P t xs 24 2 DI 18 18 15 15 12 12 9 9 6 r6 3 3 od Lo x0 9AM 10AM AM 12PM 1PM 2PM 3PM 4PM 5PM 6PM 7PM gin up Selected Latest Selected Latest Predi
16. the hub overall view two rulers are added to each compound bar The top ruler has its origin at the at hub left hub transition and helps the user in estimating how much will be left until the end of the shift The ruler below the compound bar has its origin at the left edge of the eft hub area and supports the user in estimating the total throughput for the entire shift i e what has already been done and what still remains to be done In addition to the rulers hovering the pointer over a given area brings up a text box with the exact number in a tooltip like fashion A closer look at the diagrams will also reveal that all compound bars corresponding to different service types are aligned at the at hub left hub transition This is done on purpose as this transition separates the pallets the hub no longer has to handle from those that still mean work to be done in the current shift 11 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli vera ble d E ADVANCE Live Reporter be Hub overall BB count Floor Overall All services m Left hub 2228 vumm 7 At hub 2387 2 E Declared 7975 B Scanned 443 3 z Created 405 vum i Predicted 7899 m Lefthub 1189 x Athub 1429 2 m Declared 3346 P Scanned 84 3 Created 162 vum i Predicted 3239 Overall Priority m Left hub 269 vum Em At hub 520 H 1 2 3 4 5 1
17. the order in which the bays are displayed e quasi physical order In this case the bays are displayed in the order registered in a simplified bay map The latter is simplified in the sense that it keeps the order of bays identical for both warehouses in the A B pair This allows seamless switch over to side by side comparison of the same row of bays in both warehouses L and R options in the bay set selector e al sort bays by aggregated floorspace occupation and upcoming load This measure is calculated in a multi step procedure for each bay as described later on in further detail If L or R views are selected the bays in both A and B rows are displayed in the same order In this case the highest aggregated load for the given bay in both warehouses determines the order e A sort bays by number Here bays within the same row are simply sorted by their number It is unlikely that experienced users will use this option too often and may be replaced later on with some other option deemed more useful The state of each bay is displayed by a set of bars and coloured background fields The arrange ment of these is meant to reflect physical reality in the sense that bays in the left row fill up from the right and bays in the right row fill up from the left Therefore the configuration of bay status compounds is symmetrical here the example of a left row bay will be explained 30 A D A N C E Advanced Predictive Analysis Base
18. 0 1000 15 15 14 14 13 13 12 12 10 10 9 9 8 8 Le 7 6 6 5 5 4 4 3 3 2 2 1 1 Kb 0 1000 9AM 10AM 11AM 12PM 1PM 2PM 3PM 4PM 5PM 6PM 7PM 8PM 1000 Selected Latest Selected Latest Predicted total 8489 7899 E Standard 650 2617 Predicted visible 8489 7899 E E Priority 183 3141 Actual total 1279 6226 B Special 446 469 Actual visible 1279 6226 Figure 9 During day screen for hub total Overall vum Y x1000 15 15 14 14 13 13 12 12 10 10 9 8 8 7 7 6 6 5 5 4 4 3 3 2 2 1 D 04 0 1000 9AM 10AM 11AM 12PM 1PM 2PM 3PM 4PM 5PM 6PM 7PM 8PM x1000 Selected Latest Selected Latest Predicted total 8489 7899 B Standard 650 2617 Predicted visible 4299 4009 lB Priority 183 3141 Actual total 1279 6226 B Special 446 469 Actual visible 1097 3085 Figure 10 During day screen for hub total with one of the service types Premium in our case marked invisible 17 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli vera ble Selecting a depot with the hub depot selector brings up a similar diagram with one more option as in all other screens inbound depot hub and outbound hub depot traffic are shown separately as in Fig 11 In this case a side by side view at the graphs could have been less suggestive and cumbersome to compare Therefore another set of display option buttons has been added to switch between inbound and outbound
19. 5 Gs E j Gw 0 0 50 E a kwa ww 0 75 Guss G2 yw 5 kwa ww 1 00 Gw 0j kw 9 0 kwa ww 1 25 Gk 0 Gw kaws 1 50 kw 0 kwa ww i 1 75 kwa kwa ww k s 2 00 Gw ww O waww kwa wc Figure A 1 Displayed bar length with various sensitivity settings Deliverable A D A N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Intensity of colour component 0 0 5 1 Input variable Red only Green only Red 4 green 0 0 5 1 Input variable Figure A2 Colour transitions used for bar backgrounds A 3 1 Colour transitions The green to red transition can be specified by two functions of an input t namely one for red intensity which is zero at t 0 and reaches its maximum at t 1 and another for green that has its maximum at t 0 and reaches zero at t 1 We have to keep in mind that the bars in front of the coloured areas will be white therefore the intermediate colour between green and red should not be a glaring yellow but rather a moderate brown Choosing a quadratic function for both colour intensities can fulfil all these requirements For red we use r 1 t UU 3 while for green we have g 0 6 1 t 4 39 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable Fig A2 shows the intensity curves as functions of t as well as colour samples for all between 0 and 1 For the re
20. 8 8 10 11 12 13 14 15 16 17 18 19 mH Declared 4067 EH EEN ss E 0 1 3 4 5 6 7 D 9 10 11 12 13 14 15 16 17 18 19 20 Created 243 1000 00 i Predicted 3890 Overall Special m Left hub 470 BH x00 At hub 438 3 4 5 6 z 8 9 10 11 12 13 14 15 16 17 18 19 Declared 563 D Scanned 2 0 1 x 3 4 s 6 7 8 10 1 12 13 14 15 16 7 18 19 20 Created 0 1000 x00 Predicted 770 Figure 5 Summary screen for hub overall 12 Advanced Predictive Analysis Based Decision Support Engine for Logistics AD VANCE Deliverable Switching over to a depot i e selecting a depot with the hub depot selector brings up a similar screen however with four pairs of compound bars the top one in each pair standing for the inbound depot hub and the bottom one standing for the outbound hub depot traffic as in Fig 6 In other words the top bar in the pair shows what comes in from the given depot while the bottom bar shows what the same depot takes from the hub In addition to the double bars each pair has two legends showing inbound and outbound numbers respectively CU ADVANCE Live Reporter Next days During day 1 Depot 1 Count 1 Depot 1 All services 1 Depot 1 Standard 3 4 45 x10 Outbound x10 Inbound 39 42 45 1 Depot 1 Priority 10 Outbound x10 Inbound 1 Depot 1 Special 39 4 45 10 Outbound 10 I
21. 8 worst 20 46 153X 0 1 x nominal cap Figure 16 Level 1 view covering one pair of warehouses Note that some controls in the header are hidden All services x1000 x1000 0 2 4 6 8 10 12 14 16 Figure 17 Compound progress bar in level 1 view and its recommended interpretation for esti mation of overall progress 27 A D VA N C E Advanced Predictive Analysis Based I Decision Support Engine for Logistics Deli vera ble 2 x nominal cap Figure 18 Floorspace occupation bars and their interpretation Which bar What is shown Bar length relative to Top Total warehouse occupation Total warehouse capacity Bottom Occupation of bay w largest relative load Capacity of this bay Table 2 Content of graphical display in one pair of floorspace occupation bars Each warehouse in an A B pair has one pair of floorspace occupation bars 28 Advanced Predictive Analysis Based Deliverable AD VA N CE Decision Support Engine for Logistics A Which bar Legend text Left column Middle column Right column Pallet of total Top Shed total footprints warehouse occupied capacity of bay Pallet of total Bottom Shed worst with largest footprints bay relative load occupied capacity Table 3 Content of the legend field for one pair of floorspace occupation bars Each warehouse in an A B pair has its own legend field for its own pair of floorspace occupatio
22. Project no 257398 Bd SEVENTH FRAMEWORK Co funded by the PROGRAMME European Union AD VANCE Project title Advanced predictive analysis based decision support engine for logistics Start date of project 01 10 2010 Duration 36 months End date of project 30 09 2013 7th pp topic addressed Challenge 4 Digital libraries and content ICT 2009 4 3 Intelligent Information Management Objective Small or medium scale focused research project STREP D5 3 ADVANCE Live Reporter users manual Document release V1 0 Due date of deliverable 09 2013 Document Status Final evolving on demand Abstract This document covers the functionalities of the ADVANCE Live Re porter implementation complementing the OS release of teh ADVANCE framework An appendix explaining the numerical background of display behaviour is added for readers wishing to gain a deeper understanding of the Live Reporter s behaviour A D A N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Contents Executive summary 1 Purpose and usage context of the ALR 2 Generic screens 3 Screens in hub mode 3 1 3 2 3 3 3 4 3 5 Common screen elements The Summary screen The Next days screen The During day screen Administrator functions 3 5 1 Setting business as usual scales 3 5 2 Shed layout editor 4 Screens in warehouse mode 4 1 4 2 4 3 4 4
23. as feasible we tried to provide users with options of drilling down looking at details that are present in the data but are hidden in the instant response interfaces due to their lesser immediate relevance 6 A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble Ease of use was a leading point in the design of the ALR screens and the interfaces were laid out to be as much as possible in line with the user s assumed picture of the processes Usually this is perceived as intuitive or self explanatory Nevertheless there were cases where intuitiveness was in conflict with efficient use e g explanatory text having too large a footprint in an otherwise valuable screen area or wedged in explanation obstructing easy visual orientation at first sight In such cases efficiency of use was preferred over intuitiveness we can assume that most of the time trained users will work with the ALR who will orient themselves by routine and are unlikely to ask for directions One of the keys to efficient support of routine activities is the provision of an environment that remains consistent over time as well allowing a routine of interpretation to develop and possibly serve as a common ground of correct interpretation and mutual understanding among ALR users In support of consistency we have applied the following measures wherever it was assumed useful e Same scaling for same grap
24. bserving the numbers below the graph one can see that dragging the highlighter will change the values in the selected columns to those found at the highlighted time while the atest column standing for the latest known values remains the same If the end of the time window has not yet been reached by the current time the atest column will automatically change to new values every 15 minutes as new manifest data and predictions become known A useful feature of this screen is that the contribution of various service types can be hidden or brought back again by clicking on the visibility controls to the left of the service type legend compare Figs 9 and 10 Hiding one or more service types sets the actual visible and predicted visible numbers accordingly Also hiding service types results in two more lines appearing in the graph a dotted line showing predictions summed up for the visible service types only as well as a solid line for actual total showing how much the sum of all service types would be 16 A D VA N C E Advanced Predictive Analysis Based I Decision Support Engine for Logistics De I Ivera b le Pul E ADVANCE Live Reporter E Summary Next days Hub overall MES Count Floor Overall 100
25. calculated for all services will apply to the other two Economy and Premium NextGen scales as well This scale is fixed If any bar goes beyond the scales shown it will be clipped there is no practical difference between severe overload and more than severe overload you have numbers in the legend field anyway Tick marks on the graph scales are interpreted without dimension as these are in fact footprint footprint e Bar lengths Repeat the following steps for All services Economy Premium NG Coming up Sum up all capacities for the given bay and for the given service type in all warehouses Divide the manifested scanned and entered footprints for the given service type by this sum You will get three dimensionless numbers These tell you how long the corresponding bars have to be in the top bar of your diagram n warehouse A Sum up all capacities for the given bay and for the given service type in the A warehouse Divide the at hub footprints for the given service type by this sum This will yield a dimensionless number that tells you how long the corresponding bar has to be in the in warehouse A bar of your diagram n warehouse B Sum up all capacities for the given bay and for the given service type in the B warehouse Divide the at hub footprints for the given service type by this sum This will yield a dimensionless number that tells you how long the corresponding bar has to be in the in wa
26. ches over to another screen 3 1 Common screen elements The header area of hub mode screens has some common elements that are best explained be forehand see also Fig 3 These are e The Logout and Settings buttons in the top right corner Clicking on the Logout button returns the user to the login screen while clicking on the Settings button opens a menu with setting options In the current ALR implementation the following parameters can be set by the user Business as usual scale for the quantities currently displayed separate scales can be set for hub total per depot traffic as well as each unit of measurement i e price count of separate units and occupied floorspace Shed layout and link to warehouse mode screens currently for debugging has to be consolidated in later versions e Screen selector buttons at the left side of the header bring the user to one of the hub level screens this can be interpreted as a main menu in hub mode e Display option buttons at the right side of the header control what amounts are displayed and in what unit of measurement The hub depot selector specifies if the total traffic of the hub is displayed or the inbound and outbound traffic of a selected depot Clicking on the triangle to the right of the hub depot selector opens a drop down list from which the user can select the hub or a specific depot Depots are specified by their ID number and the name of the depot e g
27. cted total 258 241 B Standard 9 54 Predicted visible 258 241 E Priority 10 117 Actual total 35 188 Bi Special 16 17 Actual visible 35 188 Figure 11 Two different During day screens can be selected for a given depot showing inbound top or outbound bottom values Note the additional inbound outbound switch below the unit selector buttons Also note that switching between these two screens preserves the vertical scaling of the graph to assist an inbound outbound comparison by stepping between these views back and forth 18 A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble Another feature available for all during day screens including the hub view is the highlighting of various actual values across the entire diagram This is meant to facilitate a graphical check of how predictions settle in towards the end of the relevant time period This option can be turned on by clicking on any of the actual values within the left side legend selected actual total latest actual total selected actual visible latest actual visible Once selected the given value is highlighted with a green background and a grey area stretches across the entire diagram up to the height corresponding to the given value Clicking on any other of the aforementioned four values changes the highlight and the graph contents in radio button fashion The highlight can be removed altogether by c
28. d I Decision Support Engine for Logistics Deli vera ble m Which warehouse At hub now Coming up Im FE Top bar is Standard Bottom bar is Priority Special Bay number Overall state of bay Figure 20 Structure of header and bay status bar in the level 2 screen e The rectangular field to the right stands for the shipments coming up Here a weighted sum is calculated separately for the Standard and for the Priority Special shipments composed of manifested but not yet tipped scanned and entered amounts Manifested amounts receive the strongest weight and as we go towards entered the weights decrease representing weaker certainty of the given shipment ending up at the hub in the current shift All calculations work with relative loads i e aggregated floorspace divided by bay capacity Two white bars signalise the amounts with their length the top bar stands for Standard while the bottom one represents Priority Special For large amounts the bars go into saturation so that they never go beyond the border of their coloured background The latter signalises the magnitude of the expected load with its colour ranging from green small to red large and black very large In line with the worst load philosophy mentioned earlier the colour is based on the maximum of Standard and Priority Special values At the moment a very simple method with fixed weights generates the colour and it is known that this is not the best
29. d to black transition we can re use equation 3 for the red component and keep green intensity at constant 0 Note that here t 1 means 100 load and t gradually decreases but never reaches zero as relative load goes further beyond 100 A 3 2 Adjusting input sensitivity Relative loads under and over the nominal 10096 mark are both transformed into the colour control variable with nonlinear functions that allow suppression or exaggeration of load status as needed Both functions have their own sensitivity parameters that will be explained in figures and equations further below Let us first consider the underfull situation i e where relative load is below 10096 If we set the relative load equal to t we would have a yellow background around 50 see Fig A 2 for an illustration Around 70 80 the colour would very much turn red even though in practice a bay filled to three quarters of its capacity would be far from critical Therefore we need a nonlinear function that keeps the colour green for most of the underfull load range and makes a faster transition to yellow and red towards the end A sensitivity parameter p 0 5 could then express which input would generate an output of t 0 5 see also Fig A 3 1 0 9 Value of sensitivity parameter p 0 8 0 7 0 75 0 85 0 6 M 0 95 0 5 0 4 0 3 Output colour control variable 0 2 0 1 0 p sj 0 01 02 03 04 05 06 07 08 09 1 Input f
30. der or switch board this is an area at the top of the screen that remains in place even if scrolling or paging is enabled for the rest of the screen Certain controls may be blanked out or pinned down to a given setting in some screens but the same controls will appear consistently at the same place once they become active The main groups of switch board controls are e Level changer this is a triangle pointing up or down allowing the user to step between levels if no specific target within the next level e g a selected bay is given In Level 1 the triangle points downward as the only choice can be Level 2 with a warehouse overview In all other screens the level changer points upward allowing the user to step back to a higher less detailed level e Shed selector this specifies which pair of warehouses the user wishes to see F or K The warehouse selector can be set in the Leve 1 screen but is pinned down in all other levels to protect the user from accidentally switching to the other pair of warehouses This mechanism is based on the assumption that one person is responsible for only one pair of warehouses on site e Bay set selector this control specifies the set of bays to be viewed in the Leve 2 screen and partly in Leve 3 as well The bay set selector appears with all options A B L and R offered for Level 2 while for Level 3 only L and R are offered e Bay order selector this sets the order in which bays are li
31. efore looking at the layout merging algorithm some definitions have to be made As shown in Fig B1 we look at the warehouse from the entrance i e from the point of view of the observation platform in the warehouse This makes left and right unambiguous Numbering of bay positions begins at the far end the exit side and each row is numbered separately Now let us take a look at Fig B2 and see how the merging algorithm works We start out with two warehouse layouts that are nearly identical but some deviations exist e Bays of the same number may occur in both rows of the same warehouse e Bays of the same number may occur in opposite rows in A and B respectively e Bays may be exclusive to one warehouse in the A B pair e Bays of the same number may be in the same row but at a different position in A and B e the order of bays in the same row may be different 51 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable Exit First position in row Left row E Right row Entry Figure B 1 Definition of positions in a warehouse We have to deal with all these phenomena while assembling a consistent layout without bay duplication while sticking to the actual physical layout as much as possible To this end a three phase algorithm is run 1 We go through both rows of the A warehouse left row first from first position to last If we encounter a new bay
32. er types and e Present the right data in the right way to the user of a given type Often it was found that current practice has little support for surveying a complex situation where the right data have to be present simultaneously side by side for direct comparison or gathering a global impression of the current situation Therefore ALR screens were designed to e Collate all available information needed for the given perspective e Present data in forms that have been assumed most adequate for the given perspective e g colours bar lengths or numbers and e Be arranged in a way that would facilitate efficient orientation for trained personnel The screens were designed with assumed observation and response times in mind i e the approximate time frame in which the user has to find relevant information in the screen and react on it Some of the screens are therefore optimised for quick response aimed at delivering relevant information with minimal interaction i e seeing everything relevant without having to click or scroll one s way through This is especially true for warehouse screens as these are used in possibly the busiest environment on site Other screens on the other hand are used in situations where more time is available to look behind the scenes and explore the development of processes for a while Here users may feel an overly simplified view as a restriction of their possibilities Therefore wherever this w
33. even new users can quickly learn those few bay numbers that are unique to one warehouse in the A B pair t is however important to understand that the capacity of a phantom bay does not add to the total capacity of all bays with the given number If for example there is a single bay of a given number in warehouse A the total capacity over all warehouses would be equal to the capacity of this single bay The use of these consolidated bay capacities is as follows Level 2 warehouse screen see also Tab A 5 reproduced here under Tab B 1 for convenience e At hub we take occupied footprints separately for each warehouse in the pair and divide these by the bay capacity in the given warehouse This is done separately for each of the two differentiated service types e Coming up since we do not know which particular warehouse the shipments will go into we divide all footprints under coming up by the sum of bay capacities over all warehouses Again this is done separately for each service type 54 A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble Level 3 warehouse screen see also Tab B 2 e Displayed range and tick marks Divide business as usual footprint by the sum of all bay capacities in all warehouses for the given bay number This will determine how many units the scale will span We have a business as usual scale for all service types only Therefore the range
34. g 14 Tue Inbound Outbound Inbound Outbound Inbound Outbound Total 345 47 219 72 380 14 265 54 437 70 278 50 E Standard 202 06 79 36 199 81 100 74 236 64 96 30 3 Priority 133 76 121 63 159 26 140 72 170 15 157 15 E Special 9 65 18 72 21 07 24 08 30 90 25 05 Figure 8 Next days screen for a selected depot Note that the outbound predictions use the same colour codes as the hub overall while inbound predictions receive a different set of colours 15 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable 3 4 The During day screen The During day screen shows the development of declared shipments and their predicted final amount for a time window preceding a hub shift The screen serves two purposes e showing the development of declarations over time and providing a sampling tool for diag nostic purposes and e giving a graphical and numerical comparison of actual declarations and predicted final amounts supporting users in assessing the quality of during day predictions made at dif ferent times of the day The screen gives the user the opportunity of comparing the latest numbers with those of an earlier point of time within the window displayed This can be done by clicking and dragging the red highlight ruler to the desired point of time note that the positioning of the highlighter follows the discretised time steps 15 minutes in our case and snaps to these points during dragging O
35. gends show progress of shipments in transit top bar with the usual colours for manifested scanned and entered shipments bay occupation in the A warehouse middle and the same for the B warehouse bottom repeated for all services Standard and Priority Special Bar lengths are displayed relative to bay capacity sum of A and B for upcoming shipments capacity in A for floorspace usage in A and the same again for B while the legend shows the same amounts in unscaled shipment floorspace All services EN Declared 49 00 BEN Scanned 39 00 Created 28 00 ES in wuA 2 00 EZ inwup 9 00 0 1 x bay cap x bay cap Figure 22 Status bars for all servigg types of a bay in a level 3 screen D V N C E Advanced Predictive Analysis Based A A Decision Support Engine for Logistics Del Ivera bl e In the Level 3 view the level changer green triangle in the top left corner points upward tapping on this control will bring the user back to Level 2 Med Live Reporter Warehouse 1 A Warehouse 2 R 5 Warehouse 1 Depot 5 All services B Declared 49 00 B Scanned 39 00 I Created 28 00 aw MiinwHA 2 00 MiinwHe 9 00 0 1 2 x bay cap x bay cap Standard Bl Declared 30 00 Bl Scanned 5 00 I Created 1 00 MiinwHA 2 00 MiinwHB 4 00 0 1 Z x bay cap x bay cap Priority Special Bl Declared 19 00 Bl Scanned 34 00 BW Created 21 00 3 in wnA 0 00 MiinwHeB 5 00 0 1 2 x bay
36. h we assume that consistent scaling for graphical display of quantities facilitates a visual judgement to develop Therefore many diagrams use a so called business as usual scale that can contain figures well within the same interval for the majority of cases Automatic stepwise scale to fit mechanisms would set in during rare exceptions only when the graph would stretch beyond the boundaries of business as usual scaling e Same meaning for same colour we keep to two fundamental colour schemes in all screens and for all user groups These colours are a discrete colour set for shipment status and inbound outbound distinction and a continuous colour chart for bay saturation in general for gradual transition from OK critical and over critical predicates of a situation now used for bay loads only e Same arrangement for same function meaning wherever possible screen layouts were designed in keeping with the same basic arrangement as well as the same order of displayed items e g various service types e Persistence over time wherever technically possible ALR screens retain their settings and scroll position if there is any until the next time they are visited This facilitates among other things a quick look at another screen without having to struggle to restore the perspective of the previous screen upon return 2 Generic screens The login screen is shared by all user groups and requests a user name and
37. he overall state dot 44 AD VANCE Advanced Predictive Analysis Based Decision Support Engine for Logistics Deliverable Divide this footprint by the sum of capacities is marked in this row A Standard A Priority B Standard B Priority Standard created Standard scanned Standard declared Priority created Priority scanned Priority declared Standard in warehouse A Standard in warehouse B Priority in warehouse A e Priority in warehouse B Table A5 Prescaling guide for the level 2 warehouse screen inputs Note that Priority actually stands for Priority NextGen A 5 Putting it all together Having given a detailed description of calculations behind the level 2 screen this section will show how all these fit together The various calculations making up the entire picture will be represented as boxes following the same general structure Fig A5 shows an elementary block representation of an operation with inputs parameters and outputs and Fig A6 shows how several elementary blocks can be assembled to a composite block Using this formalism Figs A 7 A 9 show various parts of display calculations while Fig A 10 shows how they fit together for displaying one row in the evel 2 warehouse view 45 A D VA NCE Advanced Predictive Analysis Based Decision Support Engine for L
38. il while anything over the limit would be gradually compressed either in terms of bar length or in terms of background shading We chose a stepless approach i e there is no distinct switching point at which saturation sets in but it is possible to set corner points and influence the sensitivity of the display both for bar lengths as well as colours This appendix is written for those who wish to have a detailed understanding of the mathematics behind these graphical display elements and relate the meaning of sensitivity parameters to display behaviour It should be noted that we intend to provide consistent display behaviour i e the same sensi tivity parameters to all users as this is expected to help them establish a common ground in interpreting the screen estimating the severeness of a situation in the warehouses and communi cating these to other users on site Nevertheless if there is consensus that sensitivity parameters differ notably from user expectations this appendix can serve as a guide to correct adjustment of these values A 2 Saturation in bar lengths Displayed bar lengths are a hyperbolic function of inputs which are the occupied floorspace for at hub bars or total footprint for coming up bars In both cases the inputs are in fact relative values and should be understood as a percentage of a nominal 100 value bay capacity in our case see also Tab A 5 for details The idea is that an inpu
39. le 3 for content Note that the rulers are for relative nominal capacity any bar reaching the 1 mark denotes that the given warehouse or bay is nominally full and anything beyond that means overfilled It is assumed that a bar going beyond 2x nominal capacity would signalise a severe situation and get proper attention anyway so that occupation bars are simply cropped beyond the second capacity mark The exact percentage of overfilling is however still displayed properly in the legend In the Level 1 view the level changer green triangle in the top left corner points downward tapping on this control the Level 2 screen will open 26 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics De I Ivera b le IN Q ADVANCE Live Reporter F All services x1000 BEI B Left hub 1423 0 2 4 6 8 10 12 14 16 Ban 1689 Bl Declared 5924 Bl scanned 275 created 248 I wi A total 206 5X B wH A worst 20 38 63 I wH B total 1007 23 I wn B worst 20 80 133X 0 1 x nominal cap Standard 1000 1000 Bi Left hub 701 0 2 4 6 8 10 12 14 16 Bang Declared 2297 m mmm E created 92 ES wi A total 138 3X I WHA worst 20 22 73 B wHB tota 549 12 I wi B worst 26 57 190 0 1 2 x nominal cap Priority Special x1000 x1000 Bi Left hub 722 0 2 4 6 8 10 12 14 16 Bang 686 Bl peciarea 3627 mH 3 E 2 J created 156 I wH A total 68 2 OB wh A worst 20 16 53 B wi B total 454 10 I wn
40. licking once more on the current highlight in the legend Fig 12 shows the four possible configurations with the highlight turned on 19 peiysiysiy Apese jequunu y uo Buy Aq 19u1e8303 e ponow q ued IYSYySIy sy JeA4 1u uorjoipeud y Jo pue y Aq sen eA jenjoe y Spsemo y ul 11 s suomipead Aep Zuunp Moy Bulydeyd S9 ul Joen y sdj y suit ydeus y MOjag Jequinu u A 3 y uo JuD Aq payeaijoe eq ue suesJ5s Aep Suun ul sen eA jenpe snoueA Sullu3i u3iH zT e4n2i4 Deliverable Advanced Predictive Analysis Based Decision Support Engine for Logistics AD VANCE gzz vi lqIsIA ng gzz O as ene EL EL jemeds m z loc oe 1230 Jeny EL EL jep ds m OLE og 12303 Jeny Er 99 Quot E B 8zz LOZ IqISIA Paad pl 99 Ayuoud E a gzz Loz SIGISIA P 13IP 1d Giz 0L pyepuejs m GI D 12303 Paad Dt ZOL psepueys m z ut eer 12303 payaIpaid sey penales jang pemeles iseje p np l s jseje p n l s 01x Wd Wd Wd9 WdS Wd Wat Wd Wat watt Weit wvol Wes Dir HIM Wd Wd9 WdS Wd WdE wat Wd Wdtt Weit Wett Wes Dir orx t on 3od q punoqui 6 orx Tr o L 30d q punoqu
41. n bars Left hub 1423 ES at hub 1689 ll Declared 5924 jl scanned 275 13 created 248 DY wi A total 206 5 WH A worst 20 38 63 WY wH B total 1007 23 l WH B worst 20 80 133X Figure 19 Location of various data within the legend field 29 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable 4 3 Level 2 bay overview The Level 2 screen gives a primarily graphical overview of upcoming loads and current floorspace occupation either for all bays of a warehouse or for the same row of bays in an A B pair of warehouses In this screen the F K warehouse selector is already pinned down to prevent an accidental switching over to the other set of warehouses The bay set selector buttons are all active offering the following options A all bays in warehouse A are shown B all bays in warehouse B are shown e L the left row of bays is shown for both warehouses for the A warehouse on the left and for the B warehouse on the right R the right row of bays is shown for both warehouses for the A warehouse on the left and for the B warehouse on the right The last two options L and R allow a side by side comparison of the same bays in both ware houses facilitating the established practice of filling up a bay in one warehouse while emptying its counterpart in the other one In Level 2 view the bay order selector is also active With these buttons the user can control
42. n layout at first occurrence Take note of position index for each bay added Phase 2 go through both rows of WH B append left over bays to common layout Phase 3 Subsequent sorting by position index ensures that the bay sequence is closest to reality Figure B2 Two phase merging of layouts in mirrored warehouse pairs 53 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable B 3 Handling capacities The handling of capacities needs to be addressed as these can vary from warehouse to warehouse and are handled accordingly in level 2 and level 3 warehouse screens Nevertheless some simple rules are enough to define the policies currently applied e Capacities of bays with the same number are added for all bay occurrences within the same warehouse For example the capacity of bays 92 in Fig B2 would be added and the consolidated layout for warehouse A would show one bay 92 with the aggregated capacity of all No 92 bays in warehouse A e The warehouse layout model keeps track of capacities by service type i e the capacities are added separately for both service types e f a bay is exclusive to one warehouse the consolidated layout still shows a phantom bay even in the warehouse where it is not really present The capacity is mirrored as well in order to prevent division by zero in the eve 2 warehouse diagrams This is not expected to cause confusion as
43. n the number or the status dot of the given bay 4 4 Level 3 bay details Level 3 shows details of floorspace occupation for a selected bay In fact the Level 3 screen contains a long scrollable document positioned with the viewport showing one selected bay either picked before in Level 2 or brought into view by subsequent scrolling in Level 3 The order of bays in the entire document is specified by the bay order selector its setting is preserved during switching between Level 2 and Level 3 in both directions but can also be changed by the user in any of these levels If the latter happens in Level 3 the bay in focus remains visibly in the same spot but the other bays shuffle around in the background to match the requested order Since the bay details shown are for both warehouses the options A and B are blanked out in the bay set selector The remaining options L and R specify which row of bays is listed in the long scrollable document currently displayed Changing the L R setting brings the user to the document showing the other row of bays again in both warehouses preserving the scrolling position of the row previously shown i e a change in the L R setting does not open the new list of bays at a forced first or last scrolling position The load details for each bay are shown as follows e A header identifies the bay by number warehouse F or K and name of the corresponding depot e Three sets of status bars and le
44. nbound 39 EI 45 Figure 6 Summary screen showing inbound and outbound pallets for perspective alignment remaining at the at hub left hub transition 13 x10 Outbound one depot from as ee S Bo hub Note the double bars for inbound at top and outbound at bottom and the A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable 3 3 The Next days screen The Next days screen shows predicted total throughput for e the current shift left column and e the shifts beginning on the next two workdays after the start of the current shift middle and right columns The dates shown in the screen indicate the starting day of the shift and the numbers below them denote the throughput expected for the shift beginning on the given day total and each service type Vertical scaling of the bars follows the business as usual scheme The service type names to the left of all columns serve as a legend for the colour codes as well Fig 7 shows a next days screen for the entire hub Predicted per depot data as in Fig 8 distinguish between inbound depot hub and outbound hub depot amounts and use two separate sets of colour codes The inbound colours derive from the At hub status in the Summary screen while the outbound as well as the hub total colour set is based on the basic green used for Left hub and most other interface elements Observing the Next days
45. number we append it to the appropriate row of the common layout Note that in Fig B 2 warehouse A has two bays numbered 92 in opposite rows By the time the algorithm finds the second occurrence in the right row it has already been inserted into the common layout first encounter counts and is therefore omitted Also note that the common layout records the position indices the given bays were first found at green numbers these will be needed in the final phase of the procedure 2 Now we go through both rows of warehouse B again left row first from first position to last and append those bays that have not been found in warehouse A Note that these bays are all exclusive to warehouse B otherwise they would have been already added to the common layout and skipped by the algorithm in warehouse B By the end of this phase we have a merged warehouse layout where B only bays are at the end of the row where they occur This may be still far from physical reality even a B only bay at first position would end up near the end of the row 3 In order to approximate the actual layout with the order of the merged map we have to sort the bays within the merged rows by index This is done in the last phase 52 A D A NCE Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli vera ble Warehouse A Warehouse B Common layout 23 23 92 Phase 1 go through both rows of the A WH add bays to commo
46. oes not support unambiguous determination of individual shipment footprints in a consignment see p 10 for more detail Relative floorspace 42 A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble occupation is calculated by dividing the actual floorspace occupation by the capacity of the given bay in the given warehouse for the given service type see also Tab A5 for details The bar lengths are determined using relative floorspace occupation for Standard and for Priority NextGen separately The applied formulae are 1 2 the current choice of the sensitivity parameter s the same value for both service types is given in Tab A2 The background colour behind both bars is determined by the maximum of the relative floorspace occupation values using the formulae specified in Tab A 1 Values used currently for sensitivity parameters p and q are given in Tab A 2 Parameter Current value Bar length sensitivity s 0 85 Same for both service types Underfull sensitivity p 0 85 Overfull sensitivity q 1 2 Table A2 Parameters for at hub now bar length and background colour functions A 4 2 Coming up bars The coming up bars differ technically from the at hub bars in that they e refer to the same bays in both mirrored warehouses as we do not yet know which particular warehouse the shipments will end up in e are based on a weighted sum of shipmen
47. ogistics Deli verable wn E Name of ke function oO UN CO A wn Bei en 3 O wn 2 2 Q E Figure A5 Example of a block representation with inputs parameters and outputs Name of block Something happens here Figure A6 Example of a composite block with inputs and outputs 46 Deliverable Advanced Predictive Analysis Based Decision Support Engine for Logistics ADVANCE Z j e AQ jleqap SoU 104 T Y Y 29S s sieq 1nojoo punoi3syoeq mou qny v aeq suoljejnaje gt mou qny ay Jeipeds Aquoud ui3us jeg paepueis u3ue jeg MoU qnu 3e y Jo suone n5je v euni UnuJixelA yysua jeg u33uo jeg deo Aeq oi ail 2 eds Auoug mou peo Aeg 3ndu euje3x3 deo Aeq 01 o4 paepueys mou peo Aeg 3ndu euj91x3 bd Mou qnu iy seniAnisues 2unj noJo5 10 Mou qny Wy 40 uniy su s u13u jeg 47 Deliverable Advanced Predictive Analysis Based Decision Support Engine for Logistics ADVANCE 2 e9 Q jleJap aow 104 Z p V 28S s sieq dn Suluuo3 y 104 suone noje g v 4n3 3 unuxelA Jnojoo punoj3xyoeg dn Surio aeq suoejnojeo dn 3uiuo Jelsadg Aquoud
48. ootprint relative to nominal 10096 Figure A3 Mapping of relative load to colour control variable t for loads less than nominal 100 40 A D d N C E Advanced Predictive Analysis Based I Decision Support Engine for Logistics Deli vera ble The equations for this mapping are b a i 5 t 6 where p is the sensitivity parameter set to 0 85 in the current Live Reporter implementation z is the relative load and t is the output passed on for generating the background colour Decreasing p makes the display respond sooner with a slower transition to full red while increasing p not going beyond 1 0 puts off the colour change to higher loads in turn for a more abrupt response as 100 load is approached The requirements for the overfull range load higher than 100 are somewhat different as here no theoretical upper load limit can be given At 100 load we start therefore out with t 1 0 and as the overload increases t goes gradually to 0 but never reaches it Here we can introduce a sensitivity parameter which specifies at what input the output reaches some relatively low threshold We chose 0 1 as the latter threshold as in the red to black transition see Fig A2 t 0 1 is reasonably close to black 1 0 9 oy Value of sensitivity parameter q 0 8 0 7 0 6 0 5 0 4 0 3 0 2 Output colour control variable 0 1 1 11 12 13 14 15 16 17 18 19 2 Input footprint relative to nominal
49. origin guarantees that it will depart for the hub While shipments can physically leave for the hub on the day they are declared this is not absolutely guaranteed Real life processes remain to be investi gated in this regard nevertheless we have set an expiry limit to keep declared shipments from accumulating and distorting the figures What is then displayed as declared is the cumulative amount of declared shipments between the current time and an expiry limit 4 At hub in most application cases this could actually mean known to be on the floor in one of the warehouses This is the amount of shipments warehouse personnel currently have to deal with in terms of floorspace occupation 5 Left hub this should actually mean known to be no longer on the floor Whatever protocol is applied to dispatching of shipments the hub no longer has to deal with the pallet when this state is set for the given shipment 10 A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble The states listed above should be understood as successive stages e g once a shipment is scanned it is no longer created once it is declared it is no longer scanned or created etc From another point of view a scanned declared status change of a shipment increments the number of declared shipments as displayed while registering its presence in a warehouse i e a declared at hub transition decreases the
50. ouse and bay status as well as amounts of upcoming shipments at various levels of detail e Level 1 general status of a given pair of mirrored warehouses indicating average floorspace usage as well as bays with largest current load Also the amount of upcoming traffic and progress status within the current shift are indicated e Level 2 concise view of bays within a given warehouse or side by side comparison of a row of bays in both mirrored warehouses e Level 3 detailed view of a selected bay number floorspace utilisation is shown for both warehouses Shed screens are optimised for use on a tablet device with an assumed resolution of 768x1024 i e in portrait orientation Also the screens are designed to operate even if only tapping and dragging typical user controls for touch screens are available to the user although some advanced or debugging functionalities are planned to be offered for desktop use e g pop up information via cursor hovering It should also be noted that the warehouse screens represent a simplified picture of the world mainly for two reasons e A simplified model of perfectly identical warehouse layouts allows the same bays in differ ent warehouses to be displayed side by side for quick visual comparison of their degree of usage This is meant to support the strategy of alternate filling and emptying as already practised today While the topology of the layout is constrained to be perfectly identical
51. plained in detailed text and illustrations e Fundamental terms needed by all user groups to understand the interfaces e Hub user and administrator screens e Warehouse user screens An appendix is also included explaining the mathematics behind various screens for advanced users who wish to have an in depth knowledge of the Live Reporter A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable 1 Purpose and usage context of the ALR The purpose of the ADVANCE Live Reporter ALR is to provide an interface between the core ADVANCE application and users taking part in logistics processes The ALR interface is web based i e all user groups can access the interface views specific to their tasks through a web browser although the physical device actually used may vary e g desktop computer laptop or mobile device It may be of advantage for users to familiarise themselves with the design philosophy behind the ALR screens therefore we give a brief summary of previous findings and design decisions at this point During interviews and process surveys it became clear that very often communicating data to users can present an information bottle neck while the required data are in some form already present in the system In order to improve this situation the combination of ADVANCE application plus ALR has to perform well in two aspects e Prepare data in a form useful for selected us
52. rehouse B bar of your diagram 55 AD VANCE Advanced Predictive Analysis Based Decision Support Engine for Logistics Deliverable Divide this footprint by the sum of capacities is marked in this row A Economy A Premium B Economy B Premium Economy entered Economy scanned Economy manifested Premium entered Premium scanned Premium manifested Economy in warehouse A Economy in warehouse B Premium in warehouse A Premium in warehouse B Table B 1 Prescaling guide for the level 2 warehouse screen inputs Note that Premium actually stands for Premium NextGen Divide this by the sum of what is marked in this row A Economy A Premium B Economy B Premium Overall coming up Overall in warehouse A Overall in warehouse B Economy coming up Economy in warehouse A Economy in warehouse B Premium coming up Premium in warehouse A Premium in warehouse B Table B 2 Scaling guide for the level 3 warehouse screen diagrams Note that Premium actually stands for Premium NextGen 56
53. s icon in the page header e Modification of business as usual scales for depots e Management of warehouse layout and bay capacity maps e User management 3 5 1 Setting business as usual scales Selecting the Business as usual scale option from the Settings menu opens a pop up window that allows the user to set the business as usual scale for the selected hub depot and the selected unit price count floorspace E 20 K PALLETWAYS BIRMINGHAM MES Lifts Ferme Settings dialog Business as usual scale 300 Figure 13 Business as usual scale dialogue box Care should be taken with choosing a value if the scale is set to unrealistically high values much more than the actual business as usual scale the readability of bay details in the Level 3 warehouse screens may suffer Note that this option can be removed entirely once an automatic mechanism is implemented that determines business as usual scales from historic data 3 5 2 Shed layout editor Clicking on Shed layout editor in the Settings menu opens a pop up window with the warehouse layout editor Fig 14 Here the user can add or remove bays from a given warehouse of the 21 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deliverable selected hub modify floorspace capacities in pallet footprint units or change the physical order of bays Mirrored warehouses are allowed to ha
54. screen prior to the next shift one may find the numbers of the left column fluctuating and finally settling in by the time all pallets for the day are manifested This is due to the fact that the first column or pair of columns for depot view is calculated via during day prediction which takes into account changes of the given day to gradually refine the prediction for the day In contrast all other columns or pairs of columns are calculated once every day via another day by day prediction method In the latter case gradual refinement would not make much difference anyway in view of the uncertainties for the subsequent days 14 A D A N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli vera ble MH E ADVANCE Live Reporter x Summary During day Hub overall MEZ Count Floor Overall Aug 10 Fri Aug 13 Mon Aug 14 Tue Total 12668 66 14185 86 13988 63 E Standard 5615 95 6216 71 5961 39 B Priority 6102 81 6872 11 6881 98 Bi Special 949 90 1097 03 1145 26 Figure 7 Next days screen for hub overall Note the colour codes for different service types explained in the legend CQ E ADVANCE Live Reporter x Summary During day 1h 1h 2012 08 10 20 00 1 Depot 1 Ra EE Count Ficor 1 Depot 1 Aug 10 Fri Aug 13 Mon Au
55. sted both in Leve 2 and Level 3 It is not active and hidden for Level 1 Table 1 summarises the visibility and active locked status of switch board controls in various warehouse view levels and Fig 15 shows their location within the warehouse screen header TAN s ADVANCE Live Reporter a Warehouse 1 B es sls Figure 15 Switchboard elements common to warehouse screens Some of these elements may be locked or hidden in certain warehouse screens 24 A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble L1 L2 L3 Level changer visible pointing down visible pointing up visible pointing up Shed selector visible active visible pinned visible pinned Bay set selector hidden visible all options active visible L and R active Bay order selector hidden visible active visible active Table 1 Status and visibility of switch board elements in various warehouse screen levels 4 2 Level 1 global warehouse overview The Level 1 warehouse screen gives a concise summary of one pair of mirrored warehouses The warehouse selector is active and the user can switch the view between warehouse pairs or unpaired warehouses The area below the switch board shows a warehouse pair summary e Status summary of shipments known to pass through the given pair of warehouses during the current shift or have already done so e Floorspace occupation statis
56. t of 100 would make the bar take up a certain part of the maximal bar length denoted as a sensitivity parameter s and overfill overload would be pressed into the remaining space This way the bar would theoretically display any magnitude of overload 36 A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble without going beyond the available bar space The exact mathematical relationship of input and displayed bar length is as follows 8 1 l s 1 Cx 2 y cz 1 2 where s is the sensitivity parameter x is the input taking on 1 0 for 100 and y is the bar length with 1 0 at full bar length The behaviour of the bars for different choices of s is also shown in Fig A 1 with some bar examples shown below the graph Note the vertical red line at 10096 input and the horizontal red lines showing how much of the maximal bar length is taken by the bar at this input value A 3 Generating colours The background colour of the at hub and coming up bars informs the user about the general footprint situation for both services in fact it corresponds to the maximum of Standard and Priority Special values The idea here is that the bar background would start from green totally empty gradually change to red 100 occupation and then turn to black overfilling These would be in fact two colour transitions empty to full and full to overfilled with end points
57. tics on warehouse level and for bays with the highest occupa tion percentage The above data are shown for e All service types altogether e Standard shipments occupation percentage calculated specifically for Standard bays and e Priority and Special shipments occupation percentage calculated specifically for Priority bays As it can be seen the display elements for the above three groups all services Standard and Priority Special have the same basic structure This consists of e A compound progress bar showing the progress of shipments processed by the given pair of warehouses with a legend showing exact numbers In this area everything is expressed in shipment floorspace and ruler units are placed accordingly The progress colours are the same as in the hub level summary screen except that here predictions are not displayed This part of the graph allows an estimation of overall progress green area vs orange etc which is useful when compared to the ratio of total vs remaining shift time 25 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable e Two pairs of floorspace occupation bars one for warehouse A and another for warehouse B showing current floorspace occupation relative to nominal capacity All bars are orange since these refer to shipments in at hub status Each pair contains two bars see Table 2 for content and a corresponding legend field see Tab
58. ts estimated footprints in different states of processing Also here the inputs are relative to bay capacity however since the coming up bars are meant for bays in all eligible warehouses in a hub we divide by the cumulative capacity of all suitable bays see also Tab A 5 for details Naturally service type separation is the same as with the at hub now bars so we do not add capacities across all service types As for the aforementioned weighted sum we take declared scanned and created shipments into consideration The relative footprints of these are multiplied by their specific weighting coefficients Wm Ws and We respectively and summed up Such sums are calculated separately for Standard and Priority NextGen shipments which then serve as inputs for bar length calculations Similarly to the at hub now case the maximum of these two sums is the input for background colour calculations Weighting and sensitivity parameters for coming up calculations are summarised in Tab A 3 43 A D VA N C E Advanced Predictive Analysis Based Decision Support Engine for Logistics Deli verable Parameter Current value Bar length sensitivity s 0 85 Same for both service types Underfull sensitivity p 0 85 Overfull sensitivity q 1 2 Weight for declared shipments w 0 6 Same for both service types Weight for scanned shipments w 0 3 Same for both service types Weight for created shipments w
59. ve different layouts i e the user can enter a real life layout map assuming that the bays allocated to a depot are not in disjoint sections of warehouse floorspace Nevertheless it should be noted that warehouse screens will harmonise differences between mirrored warehouses while capacities will remain different in A and B the presence and sequence of bays will be brought to a uniform order to allow side by side comparison of bays assigned to the same depot see Appendix A 5 for more detail xj Warehouse Layout Editor Hub Warehouse warnae U EE se Exil Lu u 10 B n 5 x FAITS 5 14 J 45 x x 153 5 x x 30 184 e 1 J 45 x x 30 131 76 60 x x 75 193 124 9 x x 90 60 x x 30 114 72 Te x X E e x x 45 221 x x 75 107 148 s x x 30 123 h s ts x x 30 a Entry Add left Em Add right Figure 14 Shed layout editor window and functionalities of controls 22 A D A N C E Advanced Predictive Analysis Based g Decision Support Engine for Logistics Deli vera ble 4 Screens in warehouse mode The warehouse mode screens are meant to support warehouse operations on site especially deci sions related to tipping and emptying bays The screens provide quickly recognisable information about wareh
Download Pdf Manuals
Related Search