From: Steve R. <LRA@Northcoast.com> - 2005-04-11 15:00:42
|
Sadykov Timur wrote: > Hello! > > Give me a hint, please, how could I implement in Repast a simulation model with > a complicated history-dependent behavior of agents? So that the actions of > agents at every step would depend on their previous actions. There should be > some structures which consider it; but I can't clearly see them in Repast, so > it starts seeming to me, that such a task is much easier done in a not specific > evironment, like C++ or Java, rather than in Repast? One way that people have implemented history (or memory) is to define a class that describes the actions. Then each individual records its past actions as a list of objects of that class. The individuals can then examine that list of past actions to base calculations on its history. The most recent action might be the first object on the list, etc., so the influence of past actions can be weighted by how long ago they were. Steve Railsback |