Programming for Social Scientists Lecture 6 UCLA Political Science 209-1: Programming for Social Scientists Winter 1999 Lars-Erik Cederman & Benedikt Stefansson

POL SCI Cederman / Stefansson 2 Today's topics Graphical User Interfaces (GUIs) –graphs –control panels Example 1: GraphIPD Program that endows EvolIPD with a dynamic graph and a control panel. Example 2: Elevator

POL SCI Cederman / Stefansson 3 The Graphical User Interface Objects provided to create and manage –Line graphs –Histograms –Raster images –Digraphs Data collection, calculation and updating is provided through support objects to the GUI widgets

POL SCI Cederman / Stefansson 4 GraphIPD: Adding GUI main EZGraph ModelSwarm popList Tournament newList winner ControlPanel Observer Swarm

POL SCI Cederman / Stefansson 5 Creating a graph EZGraph class sets up line graph for either –Single agent or object –Collection of agents In both cases need –Name of sequence –Type of sequence {Count, Total, Average, Min, Max} –Target (object or collection) –Name of method to get data Agent Averager Collection

POL SCI Cederman / Stefansson 6 Graph IPD: File Structure main.m Model- Swarm.m Tourna- ment.m Tourna- ment.h Model- Swarm.h Observer- Swarm.m Observer- Swarm.h Player.m Player.h

POL SCI Cederman / Stefansson 7 GraphIPD: main.m #import #import "ObserverSwarm.h" int main(int argc,const char ** argv) { ObserverSwarm * observerSwarm; initSwarm(argc, argv); observerSwarm = [ObserverSwarm create: globalZone]; [observerSwarm buildObjects]; [observerSwarm buildActions]; [observerSwarm activateIn: nil]; [observerSwarm go]; return 0; }

POL SCI Cederman / Stefansson 8 GraphIPD: ObserverSwarm : GUISwarm { int displayFrequency; id displayActions; id displaySchedule; id modelSwarm; id numGraph; } +createBegin: (id) aZone; -createEnd; -buildObjects; -buildActions; -activateIn: (id)

POL SCI Cederman / Stefansson 9 Creating an ObserverSwarm createBegin,createEnd –Initialize memory and parameters buildObjects –Build ModelSwarem –Build graphs, rasters and probes buildActions –Define order and timing of GUI events activate

POL SCI Cederman / Stefansson 10 Step I: Initializing + createBegin: aZone { ObserverSwarm *obj; obj = [super createBegin: aZone]; obj->displayFrequency = 1; return obj; }

POL SCI Cederman / Stefansson 11 Step II: Creating objects -buildObjects { [super buildObjects]; modelSwarm = [ModelSwarm createBegin: self]; modelSwarm = [modelSwarm createEnd]; [modelSwarm buildObjects]; // cont'd

POL SCI Cederman / Stefansson 12 Step II: Creating objects (cont'd) numGraph = [EZGraph createBegin: self]; [numGraph setTitle: "Number of all-C"]; [numGraph setAxisLabelsX: "Time" Y: "Number"]; numGraph = [numGraph createEnd]; [numGraph createSequence: "all-C" withFeedFrom: modelSwarm andSelector: M(getNum0)]; [numGraph createSequence: "TFT" withFeedFrom: modelSwarm andSelector: M(getNum1)]; [numGraph createSequence: "aTFT" withFeedFrom: modelSwarm andSelector: M(getNum2)]; [numGraph createSequence: "all-D" withFeedFrom: modelSwarm andSelector: M(getNum3)]; [controlPanel setStateStopped]; return self; }

POL SCI Cederman / Stefansson 13 -buildActions { [super buildActions]; displayActions = [ActionGroup create: self]; [displayActions createActionTo: modelSwarm message: M(step)]; [displayActions createActionTo: numGraph message: M(step)]; [displayActions createActionTo: actionCache message: M(doTkEvents)]; displaySchedule = [Schedule createBegin: self]; [displaySchedule setRepeatInterval: 1]; displaySchedule = [displaySchedule createEnd]; [displaySchedule at: 0 createAction: displayActions]; return self; } Step III: Building actions

POL SCI Cederman / Stefansson 14 Step IV: Activating -activateIn: (id) swarmContext { [super activateIn: swarmContext]; // If modelSwarm also has activity schedule: // [modelSwarm activateIn: self]; [displaySchedule activateIn: self]; return [self getSwarmActivity]; }

POL SCI Cederman / Stefansson 15 Major Analysis classes EZGraph –Combines several objects to gather and display data on line graph EZBin –Does same for histogram Averager, Entropy and EZDistribution –Collect statistics from collection of objects Averager EZGraph EZBin Objectbase Entropy EZ Distribution Probe

POL SCI Cederman / Stefansson 16 Using EZGraph EZGraph is a wrapper around several objects Allows creation and setup to be achieved in one fell swoop After setting title and labels must add one or more sequences Necessary steps: 1) create an instance 2) setTitle: titleString 3) setAxisLabelsX: str Y: str 4) Add sequence(s) –With feed from object –Or feed from collection: AverageSequence TotalSequence MinSequence MaxSequence

POL SCI Cederman / Stefansson 17 Creating an EZGraph A sequence with feed from collection is created with -createASequence: aString withFeedFrom: collection andSelector: M(method) where A is one of Average, Total, Min, Max or Count aString The title, made with: -setTitle: aString The axis labels, set with -setAxisLabelsX: str Y: str

POL SCI Cederman / Stefansson 18 Using EZBin EZBin, is also a wrapper around several objects This type of graph needs to be told the number of bins for histogram and interval in which values lie Necessary steps: 1) create an instance 2) setTitle: titleString 3) setCollection: collection 4) setProbedSelector:M(method) 5) setBinNum: n 6) setUpperBound: and setLowerBound:

POL SCI Cederman / Stefansson 19 Creating an EZBin Data is fed into Histogram by -setCollection: collection and -setProbedSelector:M(method) The title, made with: -setTitle: aString The axis labels, set with -setAxisLabelsX: str Y: str The number of bins is determined with -setNumBins: num and the range of values by -setUpperBound: -setLowerBound:

POL SCI Cederman / Stefansson 20 Other utility objects in Analysis Averager –Gets feed from collection and calculates: Average Total Min Max Count Entropy –Gets feed from collection and calculates entropy assuming data is probabilities EZDistribution –Subclass of EZBin, gives acces to underlying distribution of data

POL SCI Cederman / Stefansson 21 Homework Week 6 1. Modify GraphIPD sample program: a. Add a second graph plotting the average payoff of the players over time b. Replace frequency graph with histogram showing the frequency of the four strategies dynamically 2. Extend Elevator sample program to n- tenant case

POL SCI Cederman / Stefansson 22 Homework Week 6: Exercise Apartment building with elevator: What is the average waiting time?... n Two types of elevators: Type 0: "standard" Car remains where it is after use Type 1: "modified" Car returns to zero after use

POL SCI Cederman / Stefansson 23 Assignment: a. Derive theoretical waiting times for n=2. b. Generalize the program Elevator from two-tenant case to any n. c. Simulate the estimated waiting time for both elevator types. d. Plot the waiting times as a function of n. e. Which elevator type minimizes waiting? *f. Derive theoretical waiting times for n.

POL SCI Cederman / Stefansson 24 Theoretical average waiting times for elevator types 0 and 1 n = number of stories

POL SCI Cederman / Stefansson 25 Elevator: main.m int main(int argc, const char ** argv) {... for (eType = 0; eType < 2; eType++) { [tenant1 init: 1]; [tenant2 init: 2]; [elevator init: 2 type: eType]; for (repl = 0; repl < n; repl++) { if ([uniformIntRand getIntegerWithMin: 0 withMax: 1]) [tenant1 move: elevator]; else [tenant2 move: elevator]; } printf("Type: %d Time: %10.6f \n", eType, (double) ([tenant1 getTime]+[tenant2 getTime])/(double) n); } return 0; }

POL SCI Cederman / Stefansson 26 Elevator: Tenant: SwarmObject { int floor, homeFloor; int waitingTime, trips; } -init: (int) n; -setFloor: (int) f; -(BOOL)isAtHome; -(int)getTime; -move: (id)

POL SCI Cederman / Stefansson 27 Elevator: Tenant.m... -init: (int) n { homeFloor = n; if ([uniformIntRand getIntegerWithMin: 0 withMax: 1]) floor = homeFloor; else floor = 0; waitingTime = 0; trips = 0; return self; }... -move: (id) e { waitingTime = waitingTime + [e callAtFloor: floor]; if ([self isAtHome]) [e take: self toFloor: 0]; else [e take: self toFloor: homeFloor]; trips++; return self; }

POL SCI Cederman / Stefansson 28 Elevator: Elevator: SwarmObject { int floor; int type; } -init: (int) nFloors type: (int) eType; -(int)getFloor; -(int)callAtFloor: (int) f;// returns waiting time! -take: (id) t toFloor: (int)

POL SCI Cederman / Stefansson 29 Elevator: Elevator.m Elevator -init: (int) nFloors type: (int) eType { type = eType; if (type) floor = 0; else { if ([uniformIntRand getIntegerWithMin: 0 withMax: 1]) floor = 0; else floor = [uniformIntRand getIntegerWithMin: 1 withMax: nFloors]; } return self; }...

POL SCI Cederman / Stefansson 30 Elevator: Elevator.m (2)... -(int)callAtFloor: (int) f { int wait; wait = abs(floor-f); floor = f; return wait; } -take: (id) t toFloor: (int) f { [t setFloor: f]; if (type==1) floor = 0; else floor = f; return self;