EXPERIMENTAL DESIGN-BASED MULTIPLE CRITERIA OPTIMIZATION ALGORITHM TO AID MANUFACTURING DESIGN AND PROCESS ADJUSTMENT
Bryan Rosas-Matos, Mauricio Cabrera-Rios, Esmeralda Nino Perez.
University of Puerto Rico, Mayagüez Campus, Mayagüez, PR.
The aim in a multiple criteria optimization problem is to find the values of the decision variables that result in the best possible balances among all criteria in the presence of conflict. These best balances are called Pareto-efficient solutions for the most efficient result of the problem. Often times, the functions that relate the decision variables to the criteria of interest are not well established or completely known. An experimental design, which helps to build empirical approximations based on systematic sampling, is then a helpful tool in these cases. In this work, experimental design is used along with an iterative strategy to approximate the efficient frontier of the multiple criteria optimization problem. Optimality conditions are sequentially used to make this algorithm effective and precise in reaching the efficient frontier. The development of this algorithm will be carried out using injection-molding simulations where multiple criteria are considered simultaneously to effectively decide design features as well as processing conditions. At this first stage of the work, the method is presented and demonstrated through a simple-to-verify problem.