Penalty Guided Genetic Algorithm for Partner Selection Problem in Agile Manufacturing Environment

Agile manufacturing will be the future direction for the manufacturing industry in the 21
^st
century. Partner selection and risk control are important decision problems in agile manufacturing environment. The partner selection problem was investigated during the establishment of manufacturing alliance and was formulated into a nonlinear programming model. The model's objective was to maximize project success probability within the constraints of cost and time. Because of the complexity and the nonlinearity of the model, it cannot be solved by conventional methods easily. A penalty guide genetic algorithm approach was proposed in which the penalty function was adaptive and responded to the search history. Computational results from various test problems show that the algorithm efficiently and effectively searches over the promising feasible and infeasible regions to identify a final, feasible optimal, or near optimal solution.