Comparative study of genetic algorithm and simulated annealing for optimal tolerance design formulated with discrete and continuous variables

Abstract

Optimal tolerance design has been the focus of extensive research for a few decades. This has resulted in several formulations and solution algorithms for systematic tolerance design considering various aspects. Availability of different alternative manufacturing processes or machines for realization of a dimension is frequently encountered. In such cases optimal tolerance design must also consider optimal selection of a set of manufacturing processes or machines as appropriate. Such a non-linear multivariate optimal tolerance design problem results in a combinatorial and multi-modal solution space. Optimal solution of this advanced tolerance design problem is difficult using traditional optimization techniques. The problem formulation becomes more complex with simultaneous selection of design and manufacturing tolerances. The focus of the current research is on the optimal solution of this advanced and complex tolerance design problem. Genetic algorithm and simulated annealing as non-traditional global optimization techniques have been used to obtain the solution. Application of the solution techniques has been demonstrated with the help of appropriate examples. Comparison of the results establishes that the genetic algorithm is the superior of the two approaches.