Manufacturing Enterprises Procurement Cycle Optimization Algorithm Research

In manufacturing companies, optimization of the procurement cycle has become a dominant part of cost management. Making reasonable purchasing plan can effectively enhance both purchasing and inventory management, which has significant impact on the production and operation side. Purchasing cycle optimization is the key of the procurement optimization, along with the market competition changing from the cost competition to the time competition, and rapid reaction becomes the focus of supply chain management, based on time competition that requirements continuously shorten the procurement cycle to gain the competitive advantage. The procurement cycle optimization is realized by compressing the procurement time in different stages, reducing procurement funds and inventory costs.This Dissertation firstly analyzes the factors that influence the purchasing cycle including purchasing fund, the demand, raw material consume, and order transmission processing time, supplier manufacturing the goods time, delivery time and inspection time. Second it studies the basic principle of BP neural network and particle swarm optimization, advantages and disadvantages and the improvement of the commonly used methods. Based on the characteristics of the purchasing process combined with the improved particle swarm algorithm, it optimizes the neural network structure and parameter values. Lastly, according to the model basing on the improved algorithm, it achieves the procurement cycle optimizatioa The main content of this paper is as follows:1. Combined with the particle fitness value the particle swarm to linear decreasing inertia weight and dynamically adjust the learning factor to balance the global and local search ability of particle swarm algorithm, in order to improve the algorithm accuracy and convergence speed.2. The improved particle swarm optimization algorithm is used to be the BP neural network topology structure. In carrying on the optimization for the first time, the optimal number of hidden layer nodes and vector should be confirmed, and then the optimal threshold and weight of the network should be determined during the second optimization.3. Based on the analysis of a large number of historical data in manufacturing enterprises, the influence factors of purchasing cycle can be determined. Samples are selected based on the improved algorithm to study sample data and constant training to build procurement cycle optimization model, and the availability is proved by experiments.4. Design implementation procurement cycle optimization system for aluminum electrolysis enterprises. This system has realized the material classification, analysis of stock for ins and outs of raw material inventory, safety stock estimation, and purchasing material tracking and purchasing cycle prediction.