A Novel Approach of Operation Sequencing Problem in Computer Aided Process Planning

Abstract

The paper presents a novel method for finding quasi-optimal linear sequences of operations in manufacturing processes. The objective is to produce a part by minimizing the sum of machine, setup and tool change costs. The considered approach is based on a fuzzy clustering process, where the clustering algorithm divides the problem into smaller, locally solvable parts, forming disjoint groups of manufacturing tasks. Whenever the groups are already formed, then the order between their elements is established by using a kind of sorting algorithm, inserting their elements into some chains. Joining these chains according to the induced ordering of the groups we obtain a linear ordering of the tasks of the manufacturing processes. The total cost of this manufacturing process represents a good approximation of the (expected) minimal cost. Some threshold values are used in the fuzzy classification algorithm, which results a more flexible, adaptable tool for similar kind of problems. A numerical example is also presented, which follows and illustrates the mentioned theoretical background and the calculation steps. The article provides a brief overview of a possible GNU/Octave implementation of the proposed method.