Manufacturing operations outsourcing through an artificial team process algorithm

Abstract

The successful applications of artificial intelligence techniques have been widely increased during recent years. Today, outsourcing is a common strategy to enhance competitive advantages particularly for manufacturing industries. When a manufacturer receives a number of orders from customers, an internal capacity shortage is inevitable, and then the manufacturer may decide to outsource a certain set of orders. To get the joint decision on order scheduling and operation outsourcing in a parallel machine shop, an intelligent decision making technique which is based on an artificial team process algorithm (TPA) is designed in current paper. TPA is a recently evolutionary tool where operations of exploration and an epitome-based learning behavior is employed, and the operation division between elite and plain groups is elaborated. For a successful global search, the TPA lets one group (elite) explore more global spaces, while, for fast convergence, another group (plain) is assigned to local search. So, TPA is an efficient computational-based algorithm with the potential of global search during reasonable computational efforts. Additionally, a neighborhood search algorithm (NSA) with three kinds of neighborhood structures is devised to further enhance the quality of solutions at each iteration of TPA. Furthermore, the List Scheduling (LS), as a conventional scheduler, a GA and a SA are also customized for the addressed problem as benchmark algorithms. The comparisons are carried out and the results reveal the superiority of suggested method.