TrafficAssig - Traffic Assignment
Assigns the flow of the traffic network described by the NetList net using the algorithm selected by algo with the parameters given by p1, p2 and p3:
"AON" -> All or nothing
"CAPRES" -> Capres algorithm
"DSD" -> Disaggregated simplicial decomposition (only total flow is computed). p1: iterations number, p2: precision
"DSDisaggregated" -> Disaggregated simplicial decomposition. p1: iterations number, p2: precision
"FW" -> Frank-Wolfe algorithm. p1: iterations number, p2: precision
"IA" -> Incremental assignment. p1: iterations number
"IAON" -> Iterated all or nothing. p1: iterations number
"MSA" -> Method of successive averages. p1: iterations number, p2: precision
"Probit" -> Probit assignment. p1: beta, p2: precision, p3: iterations number
"ProbitE" -> MSA algorithm for the stochastic user equilibrium (MSASUE). p1:beta, p2: precision, p3: iterations number
"LogitB" -> Logit assignment (all paths used not only efficient ones used)
"LogitD" -> Logit assignment (only efficient paths are used)
"LogitMB" -> Reiterated first step logit assignment (all paths used not only efficient ones used)
"LogitMD" -> Reiterated first step logit assignment (only efficient paths are used)
"LogitBE" -> Logit equilibrium assignment (all paths used not only efficient ones used)
"LogitDE" -> Logit equilibrium assignment (only efficient paths are used)
"LogitMBE" -> Reiterated first step logit equilibrium assignment (allpaths used not only efficient ones used)
"LogitMDE" -> Reiterated first step logit equilibrium assignment (only efficient paths are used)
"NwtArc" -> Newton method bsed on nodes-links formulation (useful when there are a small number of demands)
The assigned flow, the current travel time and the disaggregated flow are stored in %net.
It is normal that for some parameters and network some methods "TrafficAssig" does not succeed to give an assignment. In particular, for Logit method, the effective domain of the stochasticity parameter is limited and depends of the network. For some networks and some methods this domain may be empty (for example LogitB with a zero travel time loop). For Logit method, the stochasticity parameter must not be too large to avoid numerical difficulties coming from the limited precision of the computation.
//
// To load an editable traffic network graph
//
%net=TrafficExample("Steenbrink");
ShowNet();
//
// The black arcs are the roads and the cyan ones are the
// traffic OD demands. We can edit the net (see the net-edition demo).
//
// To compute the traffic assignment with "DSD" algorithm
//
TrafficAssig();
ShowNet();
//
// To show the travel times select GRAPH/Display arc names/Time
//
// It is possible to assign the flow using another algorithm
// for example the incremental assignment denoted "IA"
//
%net.gp.algorithm="IA"
%net.gp.Niter=50;
TrafficAssig();
ShowNet();
//
// For Logit assignment, the stochasticity parameter
// is in the field %net.gp.theta (default 1)
//
%net.gp.algorithm='LogitMD';
%net.gp.theta=3;
TrafficAssig();
ShowNet();
//