Applying Symbiosis Mechanism and Particle Swarm Optimization on the Seller-Buyer Inventory Games

• Main Authors: Tai-Feng Tao, 陶台鳳
• Other Authors: Tsou, Ching-Shih
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/65379003678744402380

碩士 === 世新大學 === 傳播管理學研究所(含碩專班) === 94 === One of the best known problems which inventory models address is the ordering quantity problem for the buyers Traditional economic order quantity (EOQ) model and its variants lie in this category. On the opposite side, which means from the seller’s point of view, a lot of work focus on the issue of developing a pricing and replenishment scheme for the seller to minimize his total cost. The models built from the seller’s view are appropriate when a price discount is the instrument for the seller to influence buyer’s behavior in a form of compensation. No matter what view you prefer, models unilaterally built from one perspective, either buyer or seller’s view, might lead to myopic decisions. Hence, the integrated inventory control models discussing the joint optimal decisions of the seller and the buyer have been received significant attention. Above JELS models are suitable for the situation when both the seller and the buyer belong to the same organization. Although the two approaches described in the last two paragraphs are appealing, a seller-buyer inventory control problem is basically a two-person game in which both players try to maximize (or minimize) their individual gains (or costs). Analyzing the problem from buyer or seller’s perspective cannot adequately describe a competitive situation, neither could the JELS models. So, game theory, a mathematical theory dealing with decision making among multiple agents may be a more desirable approach for seller-buyer inventory systems. The challenges the decision makers confront are not only modeling the game, but also the effort devoted to optimize relevant decisions for both players. After modeling the inventory system as a Stackelberg game between the seller and the buyer, particle swarm optimization (PSO) algorithm combined with constrains handling technique based on symbiosis phenomenon in mature, are used to solve this constrained nonlinear optimization problem (CNOP). Symbiosis mechanism, which incorporates infeasible solutions into the population, makes PSO capable to obtain better exploration in the search space and, finally, finds the equilibrium result that delineates the pricing, ordering, and backordering sizes decisions for the seller and buyer.