A Bacterial Evolutionary Solution for Three Dimensional Bin Packing Problems Using Fuzzy Fitness Evaluation

摘要

This paper presents an evolutionary quasi-solution for a problem commonly occuring in practical logistics, the three dimensional version of the bin packing problem. The algorithm presented here is a variation of the bacterial evolutionary approach, and utilizes fuzzy logic in the fitness calculation. The goal is to give a useful alternative method to the basic problem, and to demonstrate that the addition of fuzzy logic elements to the fitness function increases the speed of the evolutionary process. The paper first describes the set of problems the approach applies to, then moves on to the details of every key part of the algorithm. The results from runs based on several test data sets and parameters are used to show the general effectivenessof the algorithm, and the contrast between differing penalty factors, which are used in the fuzzy evaluations of intermediate solutions. It is clearly shown that for certain sets of problems, specific values of this factor are more beneficial for the effectiveness and overall convergence of the algorithm.