Linear and non-linear assembly planning: fuzzy graph representation and GA search

This paper addresses two aspects of assembly planning: the unified representation of both linear and nonlinear assembly sequences, and the search procedure for the best sequence. Most planners introduced in research dealt with linear and nonlinear sequences separately. This paper introduces a single representation for both types by using the fuzzy graph concept and defining a fuzzy relation between product parts, thus paving the way for building general-sequence planners. At this stage, a novel approach to generate the best `feasible´ sequence is introduced. Unlike most known methods, this approach performs the search procedure on a sequence population basis rather than a part basis, by applying genetic algorithms. Consequently, it is by far less sensitive to increases in the number of product parts than other part-based-decision planners