| Moulds and Dies (M&D), mainly used for efficient mass production of industry product parts and components, are indispensably basic craft equipment for industrial production; they serve as an essential constituent for total equipment manufacturing industry. Simply put, the level of M&D manufacturing has become an important indicator for measuring a country's product manufacturing level. Accompanied by the trend that global manufacturing industry is transferring to China, Guangdong province has been gradually becoming one of most important manufacturing bases in the world with the ever-growing demand for M&D. As opposed to the average level of M&D industry of developed industrial countries, however, the level of M&D industry of our nation, especially that in the enterprise production management area, is still lagging behind.M&D manufacturing is generally in the single-project order form with resource-oriented feature. Regarded as a typical project-based operation model, the M&D production process places a very special emphasis on cross and parallelism; inside the process always exists the two-way information transmission between a project and another. M&D production management features typical requirements for resource pooling and management coordination by multiple enterprises and projects. In practical M&D production exist many uncertainties, such as task delaying, renewable resources breaking-down and new projects arriving at random time, making it always impossible for production managers to follow the original baseline schedule (or initial schedule). Only by taking effective preventive approaches and scientific dynamic scheduling and controlling strategies can make it possible to optimize the resource allocation, to eliminate the influence of uncertainties and to ensure the project execution.Supported by the National Natural Science Foundation of China (Grant No.50675039, 50875051) and the National High-Tech. R&D Program of China (Grant No.2006AA04Z132), our research was focused on uncertainty-based dynamic scheduling problem for multiple M&D manufacturing projects. From the practical M&D manufacturing situation, the description methods for uncertainty factors were herein analyzed. The following two issues have been discussed in detail as well, the decision-making mechanism of dynamic scheduling for multiple M&D manufacturing projects, and the construction, simulation and solution methods for the related scheduling models. Applications of the research achievements will be advantageous to increase production management levels of M&D corporations, and to improve M&D industry whole competitiveness of our country.The research mainly includes the following aspects.1. The manufacturing model of M&D and its characteristics were analyzed, and uncertainties and incidents impacting on multiple M&D manufacturing projects scheduling were elaborated. A predictive-reactive scheduling model for multiple M&D manufacturing projects under uncertainties was constructed.2. Corresponding to the fact that task delaying frequently happens in multiple M&D manufacturing projects process, an algorithm was proposed to solve the predictive-reactive scheduling model. A stable baseline schedule was constructed by setting time buffers based on critical chain technology, therefore, to eliminate the effect of task delaying as far as possible. According to the actual situation, the commonly used serial schedule generation scheme and parallel schedule generation scheme were modified. On this basis, an improved particle swarm optimization was proposed to solve the reactive scheduling model.3. The concepts of resource flow and resource flow network were introduced in this thesis. A method of measuring the stability of resource flow network was proposed. On this basis, a particle swarm optimization based on the priority rules was put forward to construct a stable resource flow network and to simplify the solving process of the resource flow network. Considering the frequent occurrence of tasks delaying during multiple M&D manufacturing projects execution, an algorithm was proposed to solve the predictive-reactive scheduling model with fixed resource flow constraint.4. Considering the influence of renewable resources breaking-down and task delaying on scheduling for multiple M&D manufacturing projects, an algorithm was presented to solve the predictive-reactive scheduling model. Uncertain availabilities of renewable resources were analyzed with the theory of birth-death process, and a heuristic algorithm based on two-staged optimization was proposed to solve the stable baseline schedule. According to the actual situation, the commonly used serial schedule generation scheme was further modified. On this basis, a reactive scheduling algorithm based on chaotic particle swarm optimization was proposed.5. From multi-objective optimization perspective, a multi-objective dynamic scheduling problem for multiple M&D manufacturing projects was systematically studied with considering the co-influence of such multiple uncertainties as random arrival of new projects, breaking-down of renewable resources and tasks delaying. An improved multi-objective particle swarm optimization algorithm was proposed to solve the reactive scheduling model; based on the assessment and selection method for optimal solution, the algorithm combines both the isolated points searching strategy and the elite archiving strategy.6. A system-simulation software of dynamic scheduling for multiple M&D manufacturing projects under uncertainties was developed, with which it is convenient to simulate the dynamic scheduling process by setting up different environments. Using different algorithms as well as a large variety of data, and configuring various parameters for related algorithms, the system simulation software can analyze and verify the rationality and the effectiveness of the scheduling models mentioned above; in the meantime, such software can be also of great value for systematically analyzing and comparing the performance of the related scheduling algorithms.
|
PDF Full Text Request