Managing Solution Pools with IBM ILOG CPLEX Studio

Share This Post

By watching this video, you will learn about the default pool of feasible solutions, obtaining more non-optimal solutions, setting some solution pool options, and filtering the solution pool through CPLEX. In this video, Dr. Meysam Cheramin, the Operations Research Specialist of Cresco International, will teach you how to use the CPLEX pool feature for mixed integer programming (MIP) models to evaluate and explore alternative solutions by generating more than one solution to a MIP problem. Dr. Meysam Cheramin has a Ph.D. in Systems and Industrial Engineering from the University of Arizona. He is the author of several academic papers on optimization and presented his work at several international conferences including the Institute for Operations Research and Management Sciences (INFORMS) and the Institute of Industrial and Systems Engineers (IISE). He also has a few years of experience in teaching optimization, simulation, and statistics at the university level.

Sometimes good feasible solutions are more attractive than the optimal solution or a decision maker might want to examine alternate optimal solutions. In this case, having a Pool that includes multiple optimal and feasible solutions is very useful.

The good news is that OPL supports the CPLEX pool feature for mixed integer programming (MIP) models. The pool is a way of generating and keeping more than one solution to a MIP problem. This allows you to evaluate and explore alternative solutions. For example, you might prefer a solution that also meets a secondary objective over the incumbent. By default, the pool collects the feasible solutions, if any; that is, all the solutions that worked out at the cost of not satisfying the objective by a certain gap. You can use option settings to collect more of these non-optimal solutions. If you realize after a MIP optimization that you need additional solutions, you can set the IDE to show more of them by activating the Populate solution pool option. The number of solutions you see depends on the values you select for the associated MIP/Solution pool options. These options set CPLEX to spend some extra time after the search to find more feasible solutions by using some specific strategies.

Dr. Meysam Cheramin

Meysam is an operations research specialist with 7 years of experience in optimization, simulation, and data analytics. He has been providing consulting services that leverage operations research techniques based on the idea that all solutions should focus on return on investment, profit margin improvements, and new revenue opportunities. He is keen on solving business problems using a hybrid approach of analytics, modeling, and process improvement. Meysam has built many optimization applications for various supply chain, healthcare, energy, and financial problems. Developing resilient supply chain networks for recycling permanent magnets, developing an operating room scheduling application for hospitals, and developing trade promotion and pricing optimization applications for retailers are some of his past project experiences. Moreover, Meysam has developed Cresco's Call Center Staffing Optimization application. Meysam’s technical toolbox includes Python, C++, C, CPLEX, Gurobi, FICO Xpress, GAMS, AMPL, CVX, R, SQL, SAS, MATLAB, and ARENA. Meysam holds B.S. and M.S. degrees in Industrial Engineering from the Kharazmi University and Sharif University of Technology (Tehran, Iran). He has a Ph.D. in Systems and Industrial Engineering from the University of Arizona. He is the author of several academic papers on optimization under uncertainty, supply chain optimization, and healthcare scheduling. Furthermore, he presented his work at several international conferences including the Institute for Operations Research and Management Sciences (INFORMS) and the Institute of Industrial and Systems Engineers (IISE). He also has a few years of experience in teaching optimization, simulation, and statistics at the university level.

All Posts