Inventory Control With Inventory-level-dependent Demand Rate In Supply Chain

Multi-echelon inventory control is an important issue for supply chain management in actual process of business. A reasonable and effective control policy in multi-echelon inventory system can bring the enormous economic benefits to enterprise. In theory, multi-echelon inventory control is the hotshot and difficulty of the inventory theoretical research. Therefore the study of policies for multi-echelon inventory control in supply chain has important practical value and theory meaning.This paper studies the multi-echelon inventory control problem in supply chain. Main results of this dissertation are described as the following:1. This paper discusses the development and investigative statues of the problem about inventory control. At the same time, points out the shortages of the problem research.2. Based on the researches of predecessors, this paper presents a model of series system for multi-echelon inventory control with inventory-level -dependent demand rate, which considers the phenomena that the demand rate may be influenced by the stock levels and applies the echelon concept. The model makes up shortages of the single-level inventory control and the form of demand rate considered in the model matches the actual circumstance more.3. This paper studies the generic assembly system for multi-echelon inventory control. Where, we propose a model with inventory-level-dependent demand rate, which introduces the relative weight for every stock-point to confirm the demand rate. Then, we unite the model of series system and the model of assembly system into a uniform model according to the relation between series system and assembly system.4. This paper proposes two optimization algorithms to solve the model of multi-echelon inventory control. One is genetic algorithm, which adopts feasible mixed code to avoid the penalty function, can make the solution more exact and reduce the computational requirements. The other is the relaxation-particle swarm optimization algorithm which combines serial-relaxation algorithm with particle swarm optimization. This algorithm doesn't only have the ability to global search, but also can deal with the...