Hi, I'm working on a multi-model optimization (MMO) in OptiStruct.
I'd like to solve an optimization problem with contraints defined by entire responses from multiple models.
For example, I have two input files model1.fem and model2.fem as below:
$ model1.fem DESOBJ(MIN)=1 SUBCASE 1 LABEL LINEAR STATIC ANALYSIS STATICS SPC = 1 LOAD = 2 DESSUB = 2 BEGIN BULK DESVAR,1,Thickness,1.0,1.0,5.0,, DVPREL1,1,PSHELL,1,4,,,0.0, +,1,1.0, DRESP1,1,Mass,MASS,PSHELL,,,SUM,1, DRESP1,2,dMax1,DISP,,,TY,,21, ...
$ model2.fem DESOBJ(MIN)=1 SUBCASE 1 LABEL LINEAR STATIC ANALYSIS STATICS SPC = 1 LOAD = 2 DESSUB = 2 BEGIN BULK DESVAR,1,Thickness,1.0,1.0,5.0,, DVPREL1,1,PSHELL,1,4,,,0.0, +,1,1.0, DRESP1,1,Mass,MASS,PSHELL,,,SUM,1, DRESP1,2,dMax2,DISP,,,TY,,21, ...
And MMO master input file master.fem
$ master.fem SCREEN OUT RESPRINT ASSIGN,MMO,Model1,model1.fem ASSIGN,MMO,Model2,model2.fem BEGIN BULK ENDDATA
In this case, for example, I can define a constraint by summing up the response 2 in model1.fem (dMax1) and response 2 in model2.fem (dMax2), that is
dMax1 + dMax2 >= target
In this constraint, the responses are not independent. How do I use this type of constraint? Or is it even possible?