optimization - Best algorithm for optimizing the decisions in a simulation -


i'm looking best algorithm optimise decisions made in simultaion find fast result in reasonable amount of time. simultaion number of "ticks" , occasionaly needs make decision. goal state reached. ( possible never reach goal state if make bad decisions )

there many many goal states. want find goal state least number of ticks ( tick equates second in real life." want decide decisions make goal in few seconds possible,

some points problem domain:

  • straight off bat can generate series of choices lead solution. won't optimal.
  • i have reasonable heuristic function determine decision
  • i have reasonable function determine minimum possible time cost node goal.

algorithms:

  • i need process problem 10 seconds , give best answer can.
  • i believe a* find me optimal soluton. problem decision tree large won't able calculate quick enough.
  • ida* give me first few choices in 10 seconds need path way goal.

at moment thinking start off known non optimal path goal , perhaps use simulated anealing , attempt improve on 10 seconds.

what algorithm research try solve sort of problem?

have @ limited discrepancy search, repeating increasingly loose limits on maximum discrepancy search, or beam search.

if have heuristic should able use compare individual choices - limited discrepancy search, , compare partial solutions, beam search.

see if can place upper bound on how extension of partial solution is. can prune out partial solutions can't possibly extended beat result heuristic, or best result found far in series of iterative searches increasing depth.


Comments

Popular posts from this blog

linux - Mailx and Gmail nss config dir -

c# - Is it possible to remove an existing registration from Autofac container builder? -

php - Mysql PK and FK char(36) vs int(10) -