A Mathematical Negotiation Mechanism for Distributed Procurement Problems and a Hybrid Algorithm for its Solution

TL;DR

This paper introduces a mathematical negotiation mechanism for distributed procurement in automotive supply chains, aligning supplier production with buyer costs, and proposes a hybrid PSO-A algorithm to efficiently solve the complex bi-level programming model.

Contribution

It develops a novel bi-level programming model for procurement negotiations and introduces a hybrid PSO-A algorithm to solve it more effectively than existing methods.

Findings

The hybrid PSO-A algorithm outperforms PSO-SA and PSO-Greedy in computational tests.

The mechanism aligns supplier production planning with buyer procurement costs.

The model supports partnership with valued suppliers by considering capacities.

Abstract

In this paper, a mathematical negotiation mechanism is designed to minimize the negotiators' costs in a distributed procurement problem at two echelons of an automotive supply chain. The buyer's costs are procurement cost and shortage penalty in a one-period contract. On the other hand, the suppliers intend to solve a multi-period, multi-product production planning to minimize their costs. Such a mechanism provides an alignment among suppliers' production planning and order allocation, also supports the partnership with the valued suppliers by taking suppliers' capacities into account. Such a circumstance has been modeled via bi-level programming, in which the buyer acts as a leader, and the suppliers individually appear as followers in the lower level. To solve this nonlinear bi-level programming model, a hybrid algorithm by combining the particle swarm optimization (PSO) algorithm with a heuristic algorithm based on A search is proposed. The heuristic A algorithm is embedded to solve the mixed-integer nonlinear programming (MINLP) sub-problems for each supplier according to the received variable values determined by PSO system particles (buyer's request for quotations (RFQs)). The computational analyses have shown that the proposed hybrid algorithm called PSO-A outperforms PSO-SA and PSO-Greedy algorithms.