The Research On The Multi-project Scheduling Model And Algorithm Of The Mass Customization Enterprise

With the rapid development of modern technology and the fast changing of market requirements, the traditional steady-type market has evolved into dynamic-type market. The traditional mass production mode has been challenged by the intensive market competition because it cannot satisfy the individual customers' demands. The traditional customization production mode has already been deserted for it's failure of providing low-cost and speed-to-market products. Mass customization (MC) is a new ideal for solving the pressure and paradox that enterprises are facing. It is a kind of production mode which is to organically combine the two paradoxical production modes-customization production mode and mass production mode, and produce reasonable-priced customization products with large-batch and low-cost according to the individual customers' demands.In order to satisfy the customers' demands, it's essential for the enterprises to improve the management system in every aspect, including strategic target correct localization, research and development ability promotion, cost control system consummation, human resources system promotion and so on. But, the author believed that, it's the basic strategy for the manufacture enterprise to arrange the production planning suitably and effectively to guarantee on time delivery and low cost operation.The study on the Mass Customization mainly focus on the domain of realization principle and products configuration and threr is little literature for the production plan study, the idea for the production plan is only delay strategy and modularization, the author think that this idea collides with the MC idea. The MC enterprise is a kind of pure customization enterprise, the production process of its order includes not only R&D, mould production process but also volume production process, but the traditonal methods of production plan only designed for one of the process above, for instance, the MRPII is only suitable for the volume production process and the network plan technology for the mould production plan. For realizing the core idea of MC, the process of the order should be planned and under control. According to the characteristic of the pure customization enterprise, this article carried a systemic study on its production plan.According to the project characteristic that the production operation of the mass customization enterprise has, the author believe that the mass customization enterprise has the horizontal and vertical multi-project integration environment, and then this article carried deep analysis on the environment. At the same time, based on the project management idea, a multi-project-plan and two-plane and three-stage system frame has been proposed for the mass customization enterprises to carry on the production plan and the resources disposition. Take this as the foundation, this article elaborated the contents in separately of the two planes of the system frame, which consists of rough plan of the order project and detailed plan of the multi-projects at each stage. Then the author constructed the multi-objective mathematical model of multi-project progress optimization of each stage. In order to solve the special model proposed previously, based on the current methodology of resolving the RCPSP, this article emphatically discussed the basic principle of the swarm intelligence algorithm--the particle swarm optimization algorithm and the ant colony optimization algorithm as well as the improvement research for current algorithm, and then proposed a new mixed swarm intelligence algorithm--MOAC-PSO algorithm based on the multi-object optimization and the idea of mixing the particle swarm optimization algorithm and the ant colony optimization algorithm, and analyses its application process in the multi-project progress optimization at each stage in the mass customization enterprise.Moreover, in order to confirm the validity of the model and the algorithm based on the classic instance storehouse in the project progress field, this article addresses a kind of problem instance generator-- ProGen/MMM with the user-defined function, the true significance multi-projects, multi-objectives optimization, ability to produce many kinds of problem. And developed a problem instance generator platform according to the demand of this article, the user may choose the instance type that they need, and the can set up the parameters with user-defined in the instance, the instance produced has certain randomness.Finally, based on the multi-project plan system frame, this article developed a antetype system with the main function of order project data input, transfer, inquiry, order project rough plan, various stages multi-project optimization progress at each stage and so on. At the same time, the multi-project progress optimization in various stages may be taken out from the antetype system to carry on simulated experiment for the new algorithm proposed in this article, and comparison was made with each kind of algorithm proposed by the former scholars. The experimental result indicated that the algorithm proposed in this article is effective and surpasses the meta-heuristic algorithm and the priority rule-based heuristic algorithm. This research has made several contributions:(1) Based on the analysis of the characteristic of the MC enterprise ,this article proposed the concept of horizontal multi-project integration and vertical multi-project integration for the mass customization enterprise, and constructed the multi-project-plan system frame of the mass customization enterprise;(2) Constructed a two-plane and three-stage system frame of the production plan of MC enterprise according to its characteristic;(3) Constructed the multi-objective mathematical model of multi-project progress optimization of each stage for the mass customization enterprise;(4) proposed a new mixed swarm intelligence algorithm -- MOAC-PSO algorithm;(5) Developed a problem instance generator system with high self-ajustability, the system can generate different problem instances according user's demands, these instances can be used for the performance evaluation of this article, it also can generate many kinds discrete combined optimization problems, so it resolved the bottleneck that the problem instance is not matching the model.This research provides some important train of thought and method for the mass customization enterprise to make production planning, builds a foundation to realize its short date of delivery and low cost, and makes certain contribution to the field of further project progress and production scheduling research.