topic Re: PROC OPTMODEL: Solving Optimization Problems with Hybrid Approaches_ Rob Pratt in Mathematical Optimization, Discrete-Event Simulation, and OR

PROC OPTMODEL: Solving Optimization Problems with Hybrid Approaches_ Rob Pratt

sunnyjung — Fri, 17 May 2019 08:32:16 GMT

I watched "PROC OPTMOEL: Solving Optimization Problems with Hybrid Approaches_ Rob Pratt" via youtube.

It was really helpful for me, Thank you so much.

I want to get a back data and SAS Code, Is it possible?

I am studying of Optimization to find the best location to reach everywhere easily.

I wait your reply.

Re: PROC OPTMODEL: Solving Optimization Problems with Hybrid Approaches_ Rob Pratt

RobPratt — Mon, 20 May 2019 14:56:17 GMT

Attached is the code for the p-median example.

Re: PROC OPTMODEL: Solving Optimization Problems with Hybrid Approaches_ Rob Pratt

sunnyjung — Thu, 30 May 2019 00:59:22 GMT

Thank you for your reply.

The AssignImpliesBuild is declared twice in 'pMedianMILP.sas'.

Could you explain?

Re: PROC OPTMODEL: Solving Optimization Problems with Hybrid Approaches_ Rob Pratt

RobPratt — Thu, 30 May 2019 01:33:00 GMT

The second declaration is commented out:

   /* if customer i assigned to site j, then facility must be built at j */
   con AssignImpliesBuild {i in CUSTOMERS, j in CUSTOMERS}:
      Assign[i,j] <= Build[j];
*   con AssignImpliesBuild {j in CUSTOMERS}:
      sum {i in CUSTOMERS} Assign[i,j] <= card(CUSTOMERS) * Build[j];

Both families of linear constraints enforce the logical implication that Assign[i,j] = 1 implies Build[j] = 1. The first (disaggregated) version yields a tighter linear programming relaxation, but the second (aggregated) version yields fewer constraints.