Parallel Machine Scheduling Algorithms And Software Modules Of Roller Heating Furnaces With Batching

With the fast development of the national economy, social demand for the steel products is also increasing. Therefore, the demand for the essential components of steel production-rollers is being stimulated. Roller production is an important branch of steel production. Roller productive process includes smelting, casting, heat treatment, machining and other process steps. In these processes heat treatment has a very long production cycle and high production costs, which takes a prominent position in the productive process. However, the increased demand of steel production leads to the increased amounts of bills of rollers productive enterprises, which makes the heating furnaces insufficient in amounts of heating process. For these reasons, heating process becomes the bottle-neck in roller production, which directly affects the efficiency of the whole production, also affects the output of enterprises. Therefore, it is very important that how to improve the utilization ratio of heating furnace in the heat treatment process to improve the production efficiency, reduce production costs and enhance the competitiveness of enterprises in the condition of the existing resources.This thesis was brought forward based on some national roller production enterprise and via investigating the actual situation of this enterprise, according to the relevant industry's demands for the roller production. Also, it was a constituent part of the national'973'project-the complex manufacturing process control system theory and technique basis research (2009CB320601) and the'National Natural Science Funds for Distinguished Young Scholar project'(70625001)--Optimization theory and technology.The main work of this thesis included the following sections:(1) The heating process was divided into two parts in this thesis:batching and allocating. The problem that parallel machines scheduling with fixed batching and the problem that parallel machines scheduling considering batching and allocating were studied separately.(2) Of the problem that parallel machines scheduling with fixed batching, a relevant mathematical model was established. To solve this model, a heuristic (BLPT) was designed. In the target of average of processing time, to which each furnace was planned to get close to balance the production burden. Then through large amount of experiments, the advantage of BLPT was validated through comparing to the classical heuristic LPT, SPT of parallel machines scheduling.(3) Of the problem that parallel machines scheduling considering batching and allocating, a relevant mathematical model was established. To solve this model, a genetic algorithm (GA) was designed. And a group of paratactic heuristics was also designed according to different sorting rules. The performances of these algorithms were evaluated through compare and analysis of stimulation experiments. Then to instances of different size, the suitability of the algorithms was analyzed.(4) After the design of algorithms, the heating furnaces scheduling algorithms module was accomplished using SQL Server and on the.net platen for integrated testing algorithms performances.