Intelligent approaches to tolerance allocation and manufacturing operations selection in process planning

In the modem manufacturing environment, alternative sets of manufacturing o perations can normally be generated for machining one feature of a part. Ea ch set of manufacturing operations results in a specific manufacturing cost in terms of the allocated tolerances, and requires a specific set of manuf acturing resources, such as machines, fixtures/jigs and cutting tools. In t his paper, the problems of allocating tolerances to the manufacturing opera tions and selecting exactly one representative from the alternative sets of manufacturing operations for machining one feature of the part are formula ted. The purpose is to minimize, for all the features to be machined, the s um of the costs of the selected sets of manufacturing operations and the di ssimilarities in their manufacturing resource requirements. The techniques of the genetic algorithm. and the Hopfield neural network are adopted as po ssible approaches to solve these problems. The genetic algorithm is utilize d to generate the optimal tolerance for each of the manufacturing operation s, and the Hopfield neural network is adopted to solve the manufacturing op erations selection problem. An illustrative example is given to demonstrate the efficiency of the proposed approaches. Indeed, the proposed approaches show the potential of working towards the optimal solutions to the toleran ce allocation problem and the manufacturing operations selection problem in process planning. (C) 2001 Elsevier Science B.V. All rights reserved.