MULTI-OBJECTIVE OPTIMIZATION MODELS FOR PROCESS TARGETING
Salih Duffuaa, اشرف احمد علي الجعلي
كلية علوم وهندسة الحاسب الآلي-جامعة الملك فهد للبترول والمعادن · السعودية
One of the most important decision problems in production planning and quality controlis the determination of the optimal process parameters (mean and variance). Traditionallyprocess targeting problems are formulated as a single objective optimization model. Inthis thesis the concept of multi-objective optimization is introduced to the processtargeting problem. The multi-objective models that have been developed have threeobjectives: profit maximization, income maximization and product uniformitymaximization measured by Taguchi quadratic loss function. Four multi-objectiveoptimization models are developed under four different inspection policies. The firstmulti-objective optimization model is developed under 100% error-free inspectionsystem. In the second multi-objective optimization model the inspection error freeassumption is relaxed using cut-off point for inspection instead of the originalspecification limits. The third multi-objective optimization model is developed undersampling plan with error-free inspection system. The fourth multi-objective optimizationmodel is developed where the sampling plan inspection system is subject to errors. Asuitable and reliable multi-objective optimization technique is employed to generate theset of non-inferior solutions (Pareto optimal set). The utility of the models has beendemonstrated using numerical examples. Sensitivity analysis is conducted to study theeffect of the model’s parameters and inspection errors on the sets of non-inferiorsolutions.Keywords: process targeting, quality control, 100% inspection, sampling plan, inspection error, multi-objective optimization, non-inferior solution