Home

P-Space user manual - Universitat de Barcelona

1. Defines identical features for a block of n agents Initial coordinates will of course be different for the agents in the block If they are initially IN the room then their coordinates will be assigned randomly Parameter default values are the same as the ones for the AGENT command Total number of agents specified in one or more blocks plus total number of times that AGENT is specified cannot exceed the number of agents defined with NAGENT or the maximum number of agents that is permitted NOBJECTS n Specifies the number of objects in the room They are all rectangular When graphical output is requested see OUTPUT each object is represented by a different color Maximum number of objects is 5 and minimum is 0 Default is 0 To define objects simply use as many OBJECT commands as necessary and omit NOBJECTS Nevertheless if NOBJECTS is used then it must precede any OBJECT command 0 0 El yl West East Object x1 yi x2 y2 y2 Door South x x South Figure 3 Object coordinates and door position P Space 3 3 User manual 13 OBJECT x1 y1 x2 y2 An object is defined by specifying coordinates for its northwest x1 y1 and southeast x2 y2 corners see Figure 3 Objects are areas forbidden to agents but agents can detect each other even if interposed objects exist between them Overlapped objects are not permitted Objects with corners outside of the room are trimmed off by the program Real distance between an age
2. ID12 OO OO OO oro CO OO OOO CO OO a 0 0 0 0 00 98584 21327 96100 94922 93744 92566 91388 90211 89033 87855 86677 85500 84322 83144 81967 80790 19612 78435 TAO 76082 74905 a TITZ DO RDi2 0 0 OO O 000 0000 0300 00 0 0 0 0 00 98584 IBA TY 96100 94922 93744 92566 91388 90211 89033 87855 86677 85500 84322 83144 81967 80790 a1 9612 78435 IAE 76082 74905 12129 IDO 32 P Space 3 3 User manual 33 Errors The following error messages may be emitted by P Space Error 1 Wrong number of simulations Error 2 Wrong time Error 3 IDEAL has been specified before NAGENTS Error 4 Wrong number of agents Error 5 IDEAL has been specified before AGENT Error 6 Wrong agent specification Error 7 Wrong agent initial heading Error 8 Wrong agent attention scope Error 9 Wrong agent maximu attention distance Error 10 Wrong agent mean time out Error 11 Wrong agent initial place Error 12 Wrong agent re entry door Error13 Wrong agent exit readiness Error 14 Wrong agent initial coordinate Error 15 Agent coordinates are not free or not possible Error16 IDEAL has been specified before BLOCK Error 17 Wrong block specification Error 18 Wrong block initial heading Error19 Wrong block attention scope Error 20 Wrong block maximum attention distance Error 21 Wrong block mean time out Error 22 Wrong block initial place Error 23 Wrong block re ent
3. south and west Maximum room size is x 120 y 90 and minimum is x y 10 If x lt 80 and graphical output is requested P Space 3 3 User manual 9 see OUTPUT then two additional graphical representations are shown density of occupation and a plot of variables chosen by user see SQUARE If y lt 70 then an additional plot specified with the RECTANGLE command is displayed below the room Default is x 30 y 30 Figure 1 shows a P Space sample screen P SPACE 3 3 0 Spatial behavior simulator GTICC CHASE1I Run Step End Chaos Cir 5 Euc Gray 0 500 Chao 0 000 Iner 1 000 Output CHASE1I TXT Real Ideal Distances 0 106 Figure 1 Sample screen of a P Space simulation A room with three interacting agents one object and two doors is shown in the upper left area The upper right section of the screen displays the room to scale where frequency of occupation is represented using color codes in a real screen The lower left plot shows the agents real dissatisfactions as a function of time Ideal and real distances between agents as a function of time are shown in the lower right 3x3 grid Actual graphical representations may vary depending on the room size specified and on the variables selected for the lower right see SQUARE and lower left see RECTANGLE areas P Space 3 3 User manual 10 NAGENTS n Defines the number of agents Maximum is 20 and minimum is 1 The greater the
4. BGI LITT CHR Executable English message file Spanish message file Catalan message file English help file Spanish help file Catalan help file English help navigation file Spanish help navigation file Catalan help navigation file File browser File browser English message file File browser Spanish message file File browser Catalan message file Borland DOS graphical interface file Borland graphical font file All the files must be in the same directory P Space 3 3 User manual 5 Running P Space 3 3 runs in DOS or in a Windows 98 window In order to run P Space you must first write down commands specifying the number of agents their features the room size the ideal distance models and so forth and save them as an ASCII file Commands must follow a specific P Space syntax or set of conventions described below To run the program change to the P Space directory and type one of the following commands PSPACE The program prompts for a command file name and simulates according to the commands specified in the file Default system language is used English Spanish or Catalan PSPACE i Ditto in English PSPACE e Ditto in Spanish PSPACE c Ditto in Catalan PSPACE lt file gt Simulates according to the commands specified in the file Default system language is used English Spanish or Catalan PSPACE lt file gt i Simulates according to the commands specified in the file Commands must be written in Engli
5. GRAVITY g agent ALL Defines gravity g for an agent or for all of them Agents are identified by their numbers 1 2 A single gravity may be defined for all agents by specifying ALL As many GRAVITY commands may be used as agents have been previously defined Agent numbers used in this command must refer to existing agents Minimum gravity is 0 0 and maximum is 1 0 Default is 1 0 The greater the gravity the higher the weighting of the differences between real and ideal distances on dissatisfactions The smaller the gravity the lower the weighting When gravity is high the effect of those differences on dissatisfaction depends very little on the real distances When gravity is low real distances must be small for the differences between real and ideal distances to affect dissatisfaction P Space 3 3 User manual 14 CHAOS p Specifies the probability p that the agents choose random positions within their neighborhoods at each time unit instead of those positions that minimize their dissatisfactions If p 0 then positions are never chosen randomly Default is 0 When the program is running chaos can be inserted simultaneously in all the agents by typing the hot key C A magenta vertical line is displayed on the rectangular plot see RECTANGLE indicating the time unit when the key was pressed INERTIA p Specifies the degree of inertia or constancy p 0 lt p lt 1 in the direction of movement when an agent is neutral towards
6. SAME e EXIT or NOEXIT define the agent s readiness to exit the room when it happens to step on a door threshold An agent stepping on a door threshold will exit the room if its current ideal distance to the door is 0 even if its tendency is NOEXIT it will not exit the room if its tendency is NOEXIT and its ideal distance to the door is not zero at that moment If there are no doors then this parameter is dummy Default is NOEXI T Parameters in the AGENT command can be specified progressively Initial coordinates are mandatory but the other parameters may be omitted For example we could specify initial coordinates initial heading and attention scope angle only AGENT 30 40 0 180 In this case missing parameters will be assigned their default values If the NAGENT command is used then the number of times that AGENT command is specified subsequently cannot be greater than the number of agents If the AGENT command is specified a number of times less than the number of agents then unspecified agents will be assigned their default parameter values When graphical output is requested see OUTPUT icons of 5 by 5 pixels are used for representing agents on the screen Different types of icons are available see ICON 0 0 x t 1 x t Heading Figure 2 Coordinates heading and attention scope of an agent moving around the room P Space 3 3 User manual 12 BLOCK n heading angle limit timeout IN OUT SAME RANDOM EXIT NOEXIT
7. West 7 2 South 50 Ideal distances matrix Type User All agent to agent distances are neutral All agent to object distances are neutral Age Doo Model Parameters I 2 Constant Val 0 2 1 Constant Val 0 Coordinates 1 initial coordinates are random 2 agent is out Heading 1 initial heading is random Output file Variables in output file Simul Simulation number Time Time unit or simulation step XI YIL Agent I s coordinates IDil Agent I s ideal dissatisfaction RDil Agent I s real dissatisfaction Perl Agent priority level I according to permutation AggK K th nearest neighbor standardized aggregation index P Space 3 3 Rows and columns in the TXT file correspond to timue units and variables respectively The User manual first row contains variable names are saved and values are separated with tabs Simul ime ODMDANABPWNE 1H AIAIHDUBPWNHR OO NNN NM WNHrROW pb DOOCOFRPNWWHA Uma XK N o Dil 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 RDil 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 00000 A A NS UUK WORN GO ds WB A BBR BR RRA WoOrRNW BOD NWW WWW WW WwW Ww WoOrRN WB WO
8. agents or generally a function of their interactions ideal distance models Ideal distances from agents to objects and from agents to doors can be defined as well Different models can be defined for each agent to agent agent to object and agent to door interaction As agents move to those cells in their neighborhoods where dissatisfactions tend to minimize that is to say where real and ideal distances are most similar they will keep moving around the room until their dissatisfactions are null Nevertheless it may be that due to the ideal P Space 3 3 User manual 3 distances of other agents the null level of dissatisfaction is unreachable In this case the group as a system can reach either static or dynamical stability agents stop moving or keep moving indefinitely Likewise if agents get in and out of the room or if ideal distance models that are functions of time in proximity have been defined then steady states or a dynamic equilibria can be reached and be broken periodically or chaotically P Space 3 3 Installation User manual 4 After downloading P Space copy file PSPACExx ZIP where xx indicates the program version to a directory in your system and unpack it using Pkunzip or WinZip Program P Space consists of the following files PSPACE EXE PSPENGxx TRB PSPESPxx TRB PSPCATxx TRB PSPENGxx HLP PSPESPxx HLP PSPCATxx HLP PSPESPxx NAV PSPESPxx NAV PSPCATxx NAV V EXE VENG TRB VESP TRB VCAT TRB EGAVGA
9. d2 and standard deviation abs d2 4 For example IDEAL RANDOM 5 20 will assign R a b to each non diagonal agent to agent cell independently with a gt N 5 5 4 b gt N 20 20 4 IDEAL CHASE d1 d2 Indicates that C d1 and C d2 must be assigned to all non diagonal A and B agent to agent cells respectively Agents A B B B B A B B B BB A B Agents B B B B Defaults are d1 2 d2 maxdist 4 see GEOMETRY Agents will chase each other forming a stressed circle with agent approaching agent 1 and agent n approaching agent 1 and moving away from all other agents If negative values are specified for d1 or d2 then the program assigns random values independently sampled for each cell from a normal distribution accordingly see IDEAL SUBGROUPS IDEAL CCHASE d1d2 Indicates that C d1 and C d2 must be assigned to all non diagonal A and B agent to agent cells respectively Agents A NIN B B IAJN N NIB A N Agents NI NJ B N Defaults are d1 2 d2 maxdist 4 see GEOMETRY Agents will chase each other forming a flexible circle with agent i approaching agent 1 and agent i 1 moving away from agent i and agent n approaching agent 1 and agent 1 moving away from agent n If negative values are P Space 3 3 User manual 26 specified for d7 or a2 then the program assigns random values independently sampled for each c
10. evey other agent see IDEAL If inertia equals 1 then the agent will tend to move following its initial heading if it hits a wall its heading changes If inertia equals 0 then the agent chooses a random heading at each time unit Inertia works when agents are non neutral as well although in that case it may be masked by movements caused by changes in ideal distances Inertia affects the agent decision as to which cell in the neighborhood is chosen when several cells have identical and minimum ideal dissatisfaction If inertia is 1 then that cell having the minimum difference in heading relative to the current position will be chosen If inertia is less than 1 then a probability equal to 1 p exists that a cell is randomly chosen among those sharing the minimum dissatisfaction NEIGHBORHOOD CIRCULAR MOORE NEUMANN MVONN 3 5 7 9 11 13 15 Defines the agents neighborhood a region of cells surrounding each agent at time t and limiting the area where it can move at time t 7 At time t 7 the new neighborhood surrounds the agent s new position Neighborhoods are usually symmetrical around the agent s position Size and shape of neighborhood are the same for all agents Neighborhoods of different agents can overlap Default neighborhood is CIRCULAR 3 Long neighborhood diameters cause faster agent movements and prevent that agents block each other at room corners In MOORE and MVONN neighborhoods movements can be in vertical ho
11. ignored Therefore in order to cancel out a command without deleting it just precede the command with a sign e In the specifications that follow enclose items that are optional and can be specified progressively that is ab cd means that a b c and d may be omitted or that a may be specified and b c and d omitted or that a and b can be specified and c and d omitted or that a b and c may be specified and d omitted or that a b c and d may be specified Square brackets enclose items that are mandatory A vertical bar separates mutually exclusive items NSIMULATIONS n Specifies that the process must be repeated n times Note that if you request that results be saved to a file a large n usually generates a huge file see OUTPUT Default is n 1 TIME Specifies the total number of steps or discrete time units for the simulation Total number of steps will equal t by the number of simulations Note that if you request that results be saved to a file its number of lines will equal the total number of steps at least see OUTPUT Default is t 100 ROOM x y Defines the size of the room where agents will move The room is a rectangle with horizontal side x and vertical side y where x and y are expressed in number of cells A cell is represented on the screen by a square of 5 by 5 pixels Horizontal or width and vertical or depth refer to the room location on screen Walls are called north east
12. name as the command file but extension is TXT Rows and columns in a TXT file correspond to time units and variables respectively Values are separated with tabs and thus the file can be read by SPSS StatGraphics or any spreadsheet Results saved are those specified with the VARIABLES command PERIOD or COMMA specify the punctuation sign for decimals to be used in the output file Default punctuation is PERIOD P Space 3 3 User manual 19 When OUTPUT ALL is specified the program will display graphical output and simultaneously save results in a file Format and decimal punctuation for the output file can be specified as well The program always saves a file with extension LOG containing an interpretation of the commands and a description of the results that are saved in the TXT file VARIABLES ALL COOR IDIS RDIS VIDI VLIM AGGR PERM Specifies the variables to be saved in the output file By default all variables are saved This command is dummy if no output file is requested Variables are COOR Agent coordinates IDIS Agent ideal dissatisfaction expected before moving RDIS Agent real dissatisfaction obtained after moving VIDI Variable ideal distances between agents if models with variable ideal distances have been specified VLIM Personal and social space for agents with Evolutionary models and time limits for stagnation and exile periods for agents with Hybrid models see IDEAL AGGR First to fifth neighbor aggregation
13. program assigns random values independently sampled for each cell from a normal distribution accordingly see IDEAL SUBGROUPS IDEAL COMPLEX d1 It ld d2t Indicates that D d1 lt ld d2 t and F d2 It Id d1 t must be assigned to all non diagonal A and B agent to agent cells respectively Agents A B B B B A B B BIB A B Agents B B B B A B B B B Defaults are d1 maxdist 3 It time 20 Id maxdist 4 d2 2 t time 10 where time is the total number of steps or discrete time units for the simulation see GEOMETRY Agent i is socially deprived from agent 1 and agent n from agent 1 but socially fatigated towards all other agents The group as a whole will display a global behavior of approaching and moving away with different cyclicities for different agent to agent interactions If negative values are specified for d1 It ld d2 or t then the program assigns random values independently sampled for each cell from a normal distribution accordingly see IDEAL SUBGROUPS IDEAL EVOLUTIONARY lt cf All agents generate their ideal distance to every other agent in an evolutionary way using time limit t and change factor cf in order to adapt their personal and social spaces to every other agent and as a consequence their ideal distances By default It time 20 cf 0 01 If negative values are specified for It and cf then the program assigns random values independentl
14. ELE GTICC P Space Version 3 3 User manual February 2000 Grup de Tecnologia Inform tica en Ci ncies del Comportament Computer Technology in the Behavioral Science Research Group University of Barcelona Spain P Space 3 3 User manual 1 Contents o UCU ON cio teva deletes a aaa a aeaa AARE e A a RNAAR EA Aaaa RAEL aie cca ANEA E GEEA veces 2 MSTA AUG Ri ccn dia 4 RUNNIN dl 5 O SENEE IEAI E I TS 7 Command S ya anie Etienne De ee eee Ee eee 8 O 29 A O 31 EMOS al aa at li ea ed aa AA 33 P Space 3 3 User manual 2 Introduction P Space is a program that stimulates spatial behavior in a group of agents humans robots etc Agents have attitudes towards each other which determine the constant or variable ideal distances they want to keep from other agents Agents live in a universe called room which they can exit and re enter if doors exist The room can contain objects or areas unavailable to agents The room is divided into cells and time is measured discretely A cell cannot be occupied simultaneously by more than one agent At each time unit each agent is at the center of its own neighborhood of cells an area surrounding the agent and limiting the cells where it will be able to move to at the next time unit When agents move their neighborhoods move with them An agent moves to that cell in its current neighborhood that will minimize its social dissatisfaction An agent s dissatisfaction is defined as a weighte
15. al heading is selected randomly e angle is the agent s attention scope angle measured in degrees and centered at the vector defining its current heading Default is 360 degrees e limitis the maximum distance to which the agent pays attention Default is 1000 Angle and limit define a circular sector that is the agent s attention scope The agent ignores the area within the room that is outside its current attention scope Therefore the agent acts as if its ideal distance to agents objects and doors lying in that area is neutral in Figure 2 the white area represents the agent s current attention scope e timeout is the mathematical expectancy of the amount of time that the agent will remain out of the room after exiting it or before entering it for the first time if it was initially out Actual time out is computed by sampling from an exponential distribution whose mathematical expectancy equals timeout Default is 0 e Nor OUT specify the agent s place when the simulation starts If OUT is specified then the initial coordinates are dummy If OUT is specified and the room has no doors see DOOR then the agent will remain out of the room throughout the simulation Default is N P Space 3 3 User manual 11 e SAME or RANDOM define the door used by the agent for reentering the room either the same door through which it exited the room previously or a door selected at random If there are no doors then this parameter is dummy Default is
16. ally acts as if it were neutral to agent j while their real distance is greater than critical_distance When real distance is less than critical_distance the model is activated for agent i Since that moment its ideal distance to agent j undergoes two different kinds of change A smooth change is caused by agent adapting to agent j s movements Ideal distance decreases whenever agent tries to approach agent j and receives a negative partial fustration from it that is agent gets or approaches more than predicted or whenever agent tries to leave agent J and receives a positive partial frustration from it that is it gets or leaves less than predicted Ideal distance increases whenever agent tries to leave and partial frustration is negative or when it tries to approach and partial frustration is postive If partial frustration is null ideal distance does not change If agent does not try to approach or leave agent j then ideal distance increases or decreases with probability 1 2 Smooth change in ideal distance is equal to change_factor Abrupt changes in ideal distance occur when stagnation periods of smooth changes are reached for example when ideal distance remains constant or when cyclic changes appear increase decrease increase decrease for time_limit time units In those cases ideal distance is increased by time_limit change_factor and agent iis exiled for time_limit time units When exiled id
17. d function of the discrepancy between possible future distances as estimated by the agent and ideal distances from each other agent The ideal dissatisfaction the one the agent wishes to get is its minimum possible dissatisfaction before moving and the real dissatisfaction is the one it has after moving These dissatisfactions can differ because the other agents might have moved to other cells in their respective neighborhoods and thus the agent will not reach its ideal dissatisfaction The difference between real and ideal dissatisfaction is the agent s frustration At each time unit agents decide to move within their respective neighborhoods simultaneously and independently However if two or more neighborhoods overlap then some of their cells may be candidates for more than one agent if they happen to provide minimum dissatisfactions for all of them At each time unit agents priorities to move are sorted randomly Then agents with lower priorities will be able to move only to those cells not chosen by agents with higher priorities Agents have individual features which determine their attention scopes their tendencies to exit the room their tendencies to remain outside and the effect that real distances from other agents have on the discrepancies between real and ideal distances agent gravity Ideal distances can be neutral constant increasing or decreasing as a function of time random a function of time in proximity between the
18. e time is the total number of steps or discrete time units for the simulation see GEOMETRY Agents become progressively more sociable i e they tend to approach each other as the simulation progresses If negative values are specified for d1 d2 or t then the program assigns random values independently sampled for each cell from a normal distribution accordingly see DEAL SUBGROUPS IDEAL FATIGUE d1 It Id d2 t Indicates that F d1 It Id d2 t must be assigned to all non diagonal agent to agent cells Defaults are d1 2 It time 20 Id maxdist 4 d2 maxdist 3 t time 10 where time is the P Space 3 3 User manual 27 total number of steps or discrete time units for the simulation see GEOMETRY Agents are fatigated towards each other when critical conditions for social fatigue occur If negative values are specified for d1 lt ld d2 or t then the program assigns random values independently sampled for each cell from a normal distribution accordingly see DEAL SUBGROUPS IDEAL DEPRIVATION d1 It Id d2 t Indicates that D d1 lt Id d2 t must be assigned to all non diagonal agent to agent cells Defaults are d1 maxdist 3 It time 20 Id maxdist 4 d2 2 t time 10 where time is the total number of steps or discrete time units for the simulation see GEOMETRY Agents are deprivated towards each other when critical conditions for social deprivation occur If negative values are specified for d1 lt ld d2 or t then the
19. eal distance remains unchanged When exile time is over ideal distance is P Space 3 3 User manual 23 decreased by time_limit change_factor critical_distance and the smooth change phase is resumed The time limit defining stagnation is increased by change_factor when an abrupt change occurs and the time limit defining exile is decreased by change_factor when the smooth phase is resumed This model is called hybrid because it includes two different mechanisms that cause changes in the ideal distance and also because agents behaving according to it are fatigued by stagnation and deprived by exile Example for 3 agents and 1 object Ideal user N U 1 30 100 U 1 40 500 N S 50 3 100 N Sula 30 El C 50 N N Agent 1 wants to increase its distance from agents 2 and 3 ideal distances increasing from 1 to 30 in 100 time units and from 1 to 40 in 500 time units respectively Agent 2 wants to decrease its distance from agent 1 from 50 to 3 in 100 time units while keeping a constant distance 5 from agent 3 Agent 3 wants to keep constant distances 1 and 50 from agents 1 and 2 respectively Agents 1 and 3 are neutral towards the object but agent 2 wants to keep a constant distance 30 from it General models for the group of agents as a whole can be specified They are automatically built up by the program by assigning certain ideal distance models to each agent to agent accordingly In that case the program assigns N t
20. ed by the program to slow down the process when only graphical output is requested High values of p make the process slow In old computers p must be set to low values Defining many agents and objects tends to make the program slower as well This command is dummy if no graphical output is requested see OUTPUT TRAJECTORY YES NO Specifies whether agents trajectories around the room must be displayed Default is NO This command is dummy if no graphical output is requested see OUTPUT CELLS YES NO Specifies whether the grid of cells must be displayed on the room s background Default is NO This command is dummy if no graphical output is requested see OUTPUT OUTPUT GRAPHICAL FILE ALL PERIOD COMMA Specifies the output mode If GRAPHICAL is requested then the program shows the position of the agents in the room as time progresses If room width is less than or equal to 80 then both a map showing frequency of occupation of the room cells and a plot requested with the SQUARE command are represented on the right side of the screen If room height is less than or equal to 70 then a plot requested with the RECTANGLE command is shown below the room If OUTPUT GRAPHICAL is specified then the rest of the parameters in this command are dummy OUTPUT GRAPHICAL is the default specification When OUTPUT FILE is specified no graphical output is displayed and numerical results are saved in a file Output file has the same
21. ell from a normal distribution accordingly see IDEAL SUBGROUPS IDEAL LCHASE d1 a2 Indicates that C d1 and C d2 must be assigned to all non diagonal A and B agent to agent cells respectively Agents A NIN B B IAJN N NIB JA N Agents NINI B N Defaults are d1 2 d2 maxdist 4 see GEOMETRY Agents will chase each other forming a line with agent approaching agent 1 and agent 1 moving away from agent and agent 1 moving away from agent n but agent n not approaching agent 1 If negative values are specified for d1 or a2 then the program assigns random values independently sampled for each cell from a normal distribution accordingly see IDEAL SUBGROUPS IDEAL UNSOCIABLE d1 d2t Indicates that U d1 d2 t must be assigned to all non diagonal agent to agent cells Defaults are d1 2 d2 maxdisi 4 t time 2 where time is the total number of steps or discrete time units for the simulation see GEOMETRY Agents become progressively more unsociable i e they tend to move away from each other as the simulation progresses If negative values are specified for d1 d2 or t then the program assigns random values independently sampled for each cell from a normal distribution accordingly see IDEAL SUBGROUPS IDEAL SOCIABLE d1 d2t Indicates that S d1 d2 t must be assigned to all non diagonal agent to agent cells Defaults are d1 2 d2 maxdisi 4 t time 2 wher
22. es the shortest path between them Default geometry is EUCLIDEAN The maximum possible distance in the room maxdist equals the length of the room s diagonal when geometry is EUCLIDEAN the sum of its sides when geometry is CITYBLOCK and its longer side when geometry is MAXIMUM ICON FACE CROSS CIRCLE HEADING READINESS Defines the icon used to represent agents when graphical output is requested see OUTPUT Default value is FACE When FACE CROSS or CIRCLE are specified a single type of icon is used for representing the agents both when they are inside and outside of the room and also for identifying them in other graphs see SQUARE and RECTANGLE When ORIENTATION is specified agents in the room are represented by arrows showing their approximate headings with a precision of 45 degrees while agents outside of the room are represented by arrows pointing east When TENDENCY is specified solid and empty circles are used for representing agents with EXIT and NOEXIT readinesses respectively Icons shown in the SQUARE and RECTANGLE plots are surrounded by arcs showing the areas falling out of the agent s attention scope angle When an agent is out of the room its icon is shown left to the west wall and the P Space 3 3 User manual 16 time left until the agent will reenter the room is represented by an arc surrounding the icon This command is dummy when no graphical output is requested see OUTPUT Icon colors are repeated fo
23. fault real dissatisfactions are plotted against time and a lag 10 phase plot of the mean dissatisfaction is shown In both plots only dissatisfactions between O and 0 20 are displayed nsimul 20 time room 500 80 70 nagents 6 object 30 30 40 50 gravity 0 02 all dissat 0 0 20 square phase dissat 10 link ideal chase The following commands define two agents initially in the room one object and two doors Agents are neutral to each other but both want to exit the room through a door different from the one to which they are close initially Once they exit the room they re enter it through a random door nsim time room grav agen agen obje door door 5 10000 60 60 0 5 all 310 1 360 50 10 in 50 52 1 360 50 20 in 0 15 45 60 west 7 south 50 ideal user The following commands define four agents with different gravities and attention scopes Some ideal distances are constant while some vary according to time in proximity time room 2000 120 90 block 2 1 360 block 2 1 90 gravity 0 20 1 2 gravity 0 80 3 4 P Space 3 3 User manual 30 ideal user n c 4 d 1010101 50 n c 30 n d103101031 50 n 5 1 30 10 Edd 1 30710 n c 30 fel 5 1 30 10 1 5 1 30 10 4 n e The following commands define fifteen agents with maximum attention scope All of them are initially in the room Two objects and three doors are defined Both graphical and numerical output are selec
24. fied for a non diagonal cell then the program will automatically assign Q to the diagonal cell in that same row and N to all non diagonal cells in the row Model C value It specifies that agent wants to keep a constant distance equal to value from agent object or door j P Space 3 3 User manual 21 Model R lower upper It specifies that agent wants to keep a distance from agent object or door j that varies randomly and uniformly between lower and upper at each time unit Model U start end range It specifies that agent wants to keep a distance from agent object or door j that increases linearly from start to end during the first time units of the simulation as specified by range When the ideal distance reaches end it will remain constant at that value for the rest of the simulation Model S start end range It specifies that agent wants to keep a distance from agent object or door j that decreases linearly from start to end during the first time units of the simulation as specified by range When the ideal distance reaches end it will remain constant at that value for the rest of the simulation Model F start time_limit distance_limit end range It specifies that agent wants to keep a distance from agent object or door j that will be constant and equal to start However if real distance from agent object or door j is lower than distance_limit for more than time_limit consecutive time units then ideal dista
25. hen specifying SQUARE Wrong phase plot type Wrong phase plot style Wrong RECTANGLE specification Wrong RECTANGLE type Wrong nearest neighbor number when specifying RECTANGLE Wrong agent number when specifying RECTANGLE IDEAL has been specified before NDOORS Wrong number of doors IDEAL has been specified before DOOR Wrong door wall Wrong command Could not open input file Could not open log file Could not open output file Could not open help file Could not run browser Could nat initialize graphics mode
26. ificant Finally when the agents are distributed regularly or uniformly in space it is positive and significant First to fifth nearest neighbors aggregation indexes can be requested For example RECTANGLE AGGREGATION 3 plots the first second and third nearest neighbors aggregation indexes This command is dummy if no graphical output is requested see OUTPUT or if room height is greater than 70 as it prevents the rectangular graph to be displayed DISSATISFACTION minimum maximum Defines minimum and maximum values to display when dissatisfactions are plotted on screen Default values are 0 0 and 1 0 This command is dummy if no graphical output is requested see OUTPUT DISTANCE minimum maximum Defines minimum and maximum values to display when ideal or real distances are plotted on screen Default values are 0 0 and maxdist see GEOMETRY This command is dummy if no graphical output is requested see OUTPUT FRUSTRATION minimum maximum Defines minimum and maximum values to display when frustrations are plotted on screen Default values are 1 0 and 1 0 This command is dummy if no graphical output is requested see OUTPUT P Space 3 3 User manual 18 AGGREGATION minimum maximum Defines minimum and maximum values to display when aggregation indexes are plotted on screen Default values are 3 0 and 3 0 This command is dummy if no graphical output is requested see OUTPUT PAUSE p Specifies the pause factor p gt O us
27. indexes or less if there are less than six agents defined PERM Agent permutation order IDEAL USER NEUTRAL GROUP d SUBGROUPS d1 d2 SEPARATE d CULTURAL d1 d2 RANDOM d1 d2 CHASE d1 d2 CCHASE d1 d2 LCHASE d1 d2 UNSOCIABLE d1 d2 t SOCIABLE d1 d2t FATIGUE d1 It id d2t DEPRIVATION d1 It Id d2t COMPLEX d1 Itid d2t EVOLUTIONARY t cf HYBRID cd It cf This command is mandatory and the only one that is necessary IDEAL defines a matrix of agent to agent and also agent to object and agent to door ideal distances If no objects or doors have been defined it is an n x n matrix where n is the number of agents Otherwise its dimensions are n x n b d where b and d are the number of objects and doors respectively Matrix cells contain constant or variable ideal distance models which are specified with a word and a series of numerical parameters Cell ij contains the ideal distance that agent wishes to keep from agent j if j lt n from object j if n lt j lt n b of from door j if n b lt j lt n b a If no agents have been defined number of rows in the matrix will equal 10 which is the default number of agents see NAGENTS and number of columns will equal 10 b d P Space 3 3 User manual 20 IDEAL USER Indicates that ideal distance models for all cells in the matrix will be specified by the user subsequently The matrix must be completely specified in the l
28. ines immediately following this command Ideal distances are always greater than or equal to zero When ideal distance between agents equals zero they be can eventually in contiguous cells but never in the same cell Likewise when an agent to object ideal distance is zero the agent can be eventually next to the object but never on the object However when an agent to door ideal distance is zero the real distance can be eventually zero in which case the agent will exit the room The following models are possible Model Parameters Example Quiet Q Neutral N Constant value C3 Random lower upper R110 Unsociable start end range U 130 100 Sociable start end range S 50 3 100 Fatigue start time_limit distance_limit end range F 2 20 2 50 100 Deprivation start time_limit distance_limit end range D 50 30 50 1 100 Evolutionary time_limit change_factor E 10 0 01 Hybrid critical_distance time_limit change_factor H 6 10 0 01 Only the first letter is required for specifying the model name Model Q It can be assigned to diagonal cells i i in the matrix to indicate that agent will not move during the simulation This specification will cancel any agent parameter which may affect of be related to its movement for example his tendency to exit the room see AGENT Model N It indicates that agent is indifferent to agent object or door j that is that the real distance between them will not affect agent fs dissatisfaction If Q is speci
29. n any case ideal distance is set equal to the updated personal space Initially if real distance is greater than P Space 3 3 User manual 22 social space agent acts as if it were neutral towards agent j until real distance becomes less than social space If real distance is greater than social space then a time counter is updated When the counter exceeds time_limit social space from to j increases and the counter is reset otherwise social space decreases In any case ideal distance is set equal to the updated social space If real distance is between personal and social spaces it is set equal to real distance and a time counter is updated When the counter exceeds time_limit personal and social spaces decrease and the counter is reset Whenever personal and social spaces change their increase or decrease is equal to change_factor multiplied by the neighborhood diameter If personal space increases so that it exceeds social space then the latter is increased as well If social space decreases so that it is less than personal space then social space either remains unchanged or set equal to the current personal space plus change_factor multiplied by the heighborhood diameter whichever is smaller M