Research On Technologies For Solving Large-scale Cutting Stock Problem With Mass Material And Parts

Sustainable development is the subject in current world, and the optimization of resource utilization is a detailed requirement. Manufacturing industry involves many areas, and leads to vast resource waste while turns out products and creates wealth. Optimized cutting is one of basic technologies in manufacturing industry, which is also an important ring of waste-control. As a special type of cutting stock problem (CSP), Large-scale cutting stock problem (LSCSP) traps most of the optimization algorithms available at present. As a result, large quantities of resources are wasted. In order to solve the LSCSP, this dissertation will do some research work in three different directions, including CSP with mass material (MMCSP), CSP with mass parts (MPCSP) and CSP with both mass material and parts (MMPCSP).When dealing with MMCSP, general algorithms can hardly probe into all the raw material combination methods, further more, they are lack of effective theories and technologies on material chosen, hence usually leads to fail on optimal solution on material utilization. Thus, three methods for solving the MMCSP are proposed: selection method, exchange-combination method and steeped descent method. These three methods propose hypothesis in different ways and make corresponding rules. By improving the accuracy of searching useful material combination, more efficient combinations are got and the material utilization rate is improved. The premise hypothesis, basic thinking, work flow and implementation of algorithm of the three method are expounded in the dissertation.A new class of grouping optimization method based on the cutting characteristics (CCGOM) is proposed while most of the optimization algorisms are low in time-efficiency and are likely to fall into local optimum when faced with MPCSP. The new method divide the MPCSP into several small-scale CSPs, and the combination of all results of CSPs is just the result of MPCSP. After analyzing the feasibility of CCGOM, the mathematical model of grouping optimization is established, the structure of the CCGOM was expounded, and the distinctive features including open structure, closed-loop nature, and macroscopic grouping were emphasized. Then some detailed technologies, such as the technology for deciding the number of groups, grouping strategy, parallel optimization and compensate strategy are studied. As a detailed CCGOM, grouping optimization method based on the similarity of parts (SGOM) is presented, and the work flow and some critical technologies are expounded too.To deal with MMPCSP, a systematic solution structure is analyzed, and a N-phase path control mode is proposed. The path control problem here is summarized as a 0-1 goal programming problem, whose mathematical model is established, and the solution methods are introduced in brief. Through the path control, different technologies in different aspects can harmonize well, thus a comprehensive optimal effect can be achieved.Several experiments using the steeped descent method and the SGOM are carried out to prove their validity. The experiment results indicate that the steeped descent method can find a better material combination rapidly, and can improve the material utilization rate stably. Compared with the general optimization algorithms, the SGOM is highly effective both in time-efficiency and utilization ratio of materials when solving MPCSP. Thus, to a certain extent, the validity of the technologies proposed in the dissertation is proved.At last, the technologies proposed in the dissertation were used to upgrade the optimization cutting subsystem in the"Computer Aided Design and Production Management Integrate System (CADPM)". Now, the new upgraded optimization cutting subsystem performances well and are applied in hundreds of enterprises.