Selection of Best Alternative Process Plan in Automated Manufacturing Environment: An Approach Based on Particle Swarm Optimization

Abstract

In the present flexible and automated manufacturing environment, selection of optimal process plan is a crucial decision making problem. The systematic determination of processing steps for the transformation of raw material to its finished product is identified as process planning. The real world dynamic shop floor is characterized by the availability of several machines, tools, fixtures/jigs etc., and demands the completion of several design tasks before the commencement of manufacturing actual manufacturing of a part type. Different geometrical and tolerance relationships among several features of the part types necessitate the arrangement of different setups to carry out various and hence, diverse alternative process plans to manufacture a part come into existence. Any of these feasible process plans can be used to produce the particular part type from its raw material [1], [2]. Due to the incorporation of dynamic shop floor situations such as bottleneck machines, non availability of tools, machine breakdown, etc., the process plan selection problem becomes non linear and NP hard in nature. The proliferation of Computer Aided Process Planning (CAPP) systems has made it easy and more efficient to tackle these types of non linear process planning systems. The scheduling complexity in the manufacturing systems was discussed in [2] and it was proposed that this can be reduced with the limited number of tools and auxiliary devices. The three reasons given by [2] to solve the process plan selection problem are: production cost, tool magazine capacity limitation, and reduction of auxiliary devices. Later, the process plan selection problem was attempted in [1] considering three objectives such as to minimize total time, minimize number of setups and to minimize dissimilarity among process plans. Reference [3] contributed in solving process plan selection problem using fuzzy approach to deal with the imprecise information. Reference [4] incorporated the factors such as similarity index within a process plan and degree of similarity among various process plans. They used fuzzy approach to take care of the part type processing sequence. PPS problem has also been attempted using Hybrid Hopfield Neural network and Genetic Algorithm Approach [13].