Application of Genetic Algorithm for Dimensional and Tolerance Synthesis

Abstract

Dimensional and tolerance synthesis is a key part of process planning activity that has a profound effect on the
product cost and quality. This research presents a method for synthesis of dimensions, discrete tolerances, and machine
process assignment for each machining operation of the given part(s) that minimizes the expected production cost. The
minimization problem was formulated as a non-linear mixed integer-programming problem and solved using a Genetic
Algorithm (GA), employing a non stationery penalty function. The GA coding employed generates feasible tolerances
and machine process assignment for each of the machining operations of the part(s). Optimal machining dimensions of a
part, given the associated tolerances, were determined using scaled sphere fitting method. Implementation of the
proposed method on a specific example brought out the ability of the proposed method to obtain better quality solutions
compared to conventional methods, and the potential to obtain a near optimal solution.