Genetic Algorithm for optimisation of distribution system in a single stage supply chain network with fixed charges

Abstract

This paper proposes a single stage supply chain network fixed charge transportation problem (FCTP) and a solution procedure. The objective is to find the optimum set of suppliers to be opened and determine the quantities to be supplied, so that the total cost of distribution can be minimised. The problem chosen goes beyond the traditional mathematical programming and it becomes a non-polynomial (NP) hard problem, particularly while considering the fixed charges. Genetic Algorithm (GA) approach is proposed in this paper for solving. GA is applied for optimisation of the distribution network to provide optimal or near optimal solution. To demonstrate the efficiency of the proposed method in FCTP, GA is compared with spanning tree-based GA with the aid of numerical examples. Following application and analysis of this methodology, the proposed method is identified to be efficient concerning the total distribution cost. Further, the accuracy and procedure of selecting the suitable suppliers for meeting the customer needs are found simple and systematic in approach.

Research and Application of Supply Chain Distribution Optimization Model based on Improved Genetic Algorithm

“…Select V chromosomes divided after selecting break points from the same set of code string, select randomly for V chromosomes and make crossover operation. Select chromosomes randomly, select two gene points randomly from gene site to make mutation operation and turn the genes between two gene points [12][13].…”

“…Y Y , code the demand points based on distance D IJ from small to big, take distribution center as 0 point, number the clients points from 1 to 40 and code in way of real number coding [14][15][16][17].…”

“…Deal with illegal chromosome through pretreatment operator. Find two endpoints 10 and 17 of overload path on overload path D bk3 = [10,13,18,8,9,1,5,17] and adjust chromosome through moving breakpoint location. After finding out endpoints 10 and 17, determine to move the 10 point to the left or move the 17 point to the right through comparing the vehicle load capacity of the path D bk3 = [14,15] on the left of 10 and path D bk3 = [11,7,6] on the right of 17.…”

“…It can be found out that the initial population includes a lot of tour paths, while this result in practical distribution is not reasonable and not realistic in distribution activities. For example, in K initial 1 = [9,3,9,11,7,17,14,8,9,3,3,1,2,4,1,13,1,17,11,16], path [9 3 9] is a tour path, which is not allowed in operation. Design pretreatment operator and deal with the repeated path and tour path in the path.…”

Glowworm swarm optimisation algorithm for nonlinear fixed charge transportation problem in a single stage supply chain network

“…Hajiaghaei-Keshteli (2011) developed GA and artificial immune algorithm for selecting some potential places as distribution centres in order to supply demands of all customers. Manimaran et al (2011) developed a binary coded GA for single stage SCN with fixed costs.…”