Trials not treating values correctly
Heya, I'm a long-time Simul8 user with a bit of a love-hate relationship with the software.
I apologise for not being able to go into exact details with my model due to its commercial sensitivity, but if anybody happens to have any insight into my current problem I would be very grateful
I'm currently working with a deterministic model and trying to use OptQuest for the first time to find an optimal set of resource availability; I believe OptQuest would be working just fine if it wasn't for the fact that it is required to run trials, and it is how trials are behaving that are my problem.
The Results Manager currently contains a global variable that's being used as OptQuest's objective, and 27 variables controlling availability of the resources I would like to find an optimal solution for, held as decision variables. The issue I have is that whenever a trial is run, the results for both the objective and decision values are given as [Expected_Result / Runs_In_Trial]. For example, if I run a trial with 3 runs with a resource's availability set to 5, Results Manager will tell me the average availability was 1.67.
While I'm not certain this is true, I'm strongly suspecting resources in my example are running with a value of 5 for one run, and a value of 0 for all others.
I have tried the following:
-> Using Resource.Number Available as the decision variables.
-> Using global variables as decision variables, and setting each Resource.Max Available value to the global value in On Reset logic.
-> Using global variables as decision variables, and setting each Resource.Max Available value to the global value in On Start Run logic.
-> Calculating the objective variable during On End Run logic.
-> Calculating the objective variable within a block of VL that's triggered immediately before the end of a run.
All of these methods result in the same problem.
If anybody has encountered a similar problem themselves, have any ideas about what may be causing it, or suggestions of other things to try then I will be very grateful. Thanks in advance!