Investigating supply chain performance under game theory framework using intelligent particle swarm optimisation. (2016)

• Record Type: Journal Article
• Authors: Tyagi, Annu; Tyagi, Satish
• Is Part Of: International journal of business performance and supply chain modelling. Volume 8:Number 3(2016)
• Page Start: 201
• Page End: 221
• Publication Date: 2016
• Subjects: game theory -- normal distribution -- demand allocation -- particle swarm optimisation -- supply chain performance -- intelligent PSO -- supply chain management -- SCM -- mathematical modelling -- partner profits -- transport costs -- facility utilisation -- multi-objective optimisation; Business logistics -- Research -- Periodicals; 658.5005
• Languages: English
• ISSNs: 1758-9401

Abstract

Game theory has been extensively used for analysis of situations comprising of multi-agents and their strategies. Supply chain can be defined as a network of multi-organisations interacting with each other during the decision analysis. Therefore, this paper exploits the salient features of game theory in mathematically modelling a supply chain problem and investigating its functioning under various alliances among partners of the same stage. The proposed structure of supply chain considers four different stages in the illustrative example. Profit of an individual partner at each stage while satisfying the constraints is considered in the objective function. In addition, transportation cost and facility utilisation within the whole supply chain are also targeted. Normalised values of different objectives are combined to formulate a multi-objective optimisation problem. This paper introduces a novel intelligent particle swarm optimisation algorithm which is embedded with two beneficial attributes viz.: 1) normal distribution in traditional particle swarm optimisation; 2) time varying acceleration coefficients. The computational experiment finds that maximum profit is gained when players are in union. It is also evident from results that the proposed algorithm outperforms over other variants of algorithm for the underlying problem thereby authenticating its superiority.