Optimizing Supply Chain Management using Gravitational Search Algorithm and Multi Agent System

TL;DR

This paper presents a novel approach combining a modified gravitational search algorithm with a multi-agent system to optimize demand-supply management in grain distribution, enhancing adaptability and reducing communication bottlenecks.

Contribution

It introduces a new integrated model using GSA and MAS for supply chain optimization, specifically applied to food grain distribution, improving system responsiveness and efficiency.

Findings

Effective demand-supply balancing demonstrated in grain distribution

Reduced communication overhead in supply chain management

Enhanced system adaptability to environmental changes

Abstract

Supply chain management is a very dynamic operation research problem where one has to quickly adapt according to the changes perceived in environment in order to maximize the benefit or minimize the loss. Therefore we require a system which changes as per the changing requirements. Multi agent system technology in recent times has emerged as a possible way of efficient solution implementation for many such complex problems. Our research here focuses on building a Multi Agent System (MAS), which implements a modified version of Gravitational Search swarm intelligence Algorithm (GSA) to find out an optimal strategy in managing the demand supply chain. We target the grains distribution system among various centers of Food Corporation of India (FCI) as application domain. We assume centers with larger stocks as objects of greater mass and vice versa. Applying Newtonian law of gravity as suggested in GSA, larger objects attract objects of smaller mass towards itself, creating a virtual grain supply source. As heavier object sheds its mass by supplying some to the one in demand, it loses its gravitational pull and thus keeps the whole system of supply chain per-fectly in balance. The multi agent system helps in continuous updation of the whole system with the help of autonomous agents which react to the change in environment and act accordingly. This model also reduces the communication bottleneck to greater extents.