I want to create a multi-agent simulation model for a real word production process to evaluate some dispatch rules. Simulations should create event logs to evaluate the timing effect of dispatch rules compared to real production event logs.
How can I incorporate βcurrent simulation timeβ into such a multi-agent intensive messaging simulator?
Background:
Classical discrete event modeling (which tracks time progress well) cannot be applied here, since agents in the system present relatively complex behavior and routing requirements, and dispatch rules require them to communicate frequently. This and other complexities of the process also preclude a central planning approach.
In the science of production, there are thousands of documents that use multi-agent modeling to solve some of the problems associated with production. However, I have not yet found a document that describes in detail the internal work or implementation details of these simulations in the required detail.
Unfortunately, using the shortest process time for a discrete time step in the system may not be feasible, since the process time range is from 0.1 to 24 hours. There is a possibility that my simulation will be used for if-if evaluations in the project later, so the simulation should be performed as quickly as possible - there is no way to run the simulation overnight.
The size of the problem is about 500 resources and 1,000 to 10,000 product agents, most of which are completed and are not involved in any additional relationships or resources.
, - , " ". , , . .
.
, - , . . , , .
, , - DES Agent . , , : http://code.google.com/p/tidra-framework/, .