Originally posted by: SystemAdmin

I want to decompose a complex stochastic network flow problem (essentially a MILP) into several subproblems by conducting Lagrangian relaxation on some constraints. The problem is then decomposable, say by products and I want to solve the smaller subproblems consecutively. In AMPL one can easily define an indexed set of objective functions, e.g.

minimize Total_Cost{p in PRODUCTS}: total_cost[p];

and solve iteratively by looping over p.

Is there a similar, elegant way to accomplish this in OPL? As of now, I fear I have to provide individual model and data files for each of the p decomposed submodels.
#DecisionOptimization
#OPLusingCPLEXOptimizer