Enhancing Procurement Efficiency of Collaborative Supply Networks Using Multi-Agent Actor- Critic Algorithm

Organizations strive for profitability while balancing environmental and social factors, with the procurement function playing a pivotal role. The key goal for supply organizations is to reduce the cost of products while maintaining other performance metrics. This paper introduces a novel approach to negotiating costs through iterative procurement auctions, leveraging multi-agent actor-critic algorithms. Our study focuses on buyers, financial and risk impact assessors, and other key entities within the procurement ecosystem, each equipped with actor and critic components. The actor component takes the decisions, while the critic component evaluates choices based on financial metrics. The primary objective is to source products with minimal cost so that the profit margin is improved. This multi-agent system forms an adaptive decision-making framework that learns from historical data and dynamically interacts within the procurement environment. By decentralizing decision-making, the collaborative efforts of individual agents contribute to the overarching goal of minimizing cost. The results show that through the proposed iterative process, the procurement efficiency is increased, and the system meets the goals with lower costs while satisfying other performance metrics. This innovative approach minimizes the need for extensive input from procurement managers, presenting a promising advancement in optimizing the intricate landscape of procurement decision-making.