Forward and Backward Planning Approaches for Process Sequencing in CAPP

摘要

Process sequencing is a very important part of process planning task. Several considerations are involved in this task. One of most important considerations is feature precedence network (FPN) and constraints that the network imposes on the process sequence. Furthermore, the point from which sequencing starts carries great importance because of process clustering. This paper addresses issues of forward and backward planning in the process sequencing task of automated process planning. Algorithms for network optimization (Dijkstra and Space search) are explained and various heuristic functions discussed. The procedures for feature process sequencing and process cluster sequencing (simultaneous process clustering and process sequencing) are described for both forward and backward planning. Possibilities for different process clusters, which are due to the fact that FPN does not enforce total order for machining features, are shown. Forward planning algorithm from stock and backward planning algorithm from the final part with Dijkstra optimization are explained. Application of these algorithms is shown on several examples. By providing insight into very important task of process sequencing, the paper provides criteria for selection of proper approach for particular mechanical part in process planning practice and for implementation in CAPP systems.