Modeling the fuzzy cold storage problem and its solution by a discrete firefly algorithm

摘要

Maintaining an appropriate environmental condition plays a vital role in preventing food poisoning and food spoilage. Thus, how to allocate a large number of perishable food products with different intrinsic characteristics to a cold store with different cabins is quite important for decision makers. In this paper, we define a mathematical model of the fuzzy cold storage problem (FCSP), which can be regarded as a complex variant of the knapsack problem. Profits of perishable food products are fuzzified and represented by triangular fuzzy numbers (TFNs). To solve the fuzzy system, we use the k preference integration as the defuzzification method. Given that the FCSP is highly combinatorial, we design a new discrete optimization algorithm which combines the firefly algorithm (FA) and the greedy algorithm to get near-optimal solutions. In this algorithm, we use the Hamming distance to denote the distance between two fireflies and propose a four-phase repair operator to correct and optimize the solutions. Furthermore, processes of initialization and movement of the brightest firefly are also improved to strengthen the algorithm. Finally, computational simulations with randomly generated data are analyzed and reported to evaluate the performance of the proposed algorithm.