Fonotov A. Automatic system of guillotine cutting on bases genetic programming (on example furniture manufac-ture)

The Subject of inquiry is technological process of the saw timber, woodchip boards and fibreboard guillotine cutting. The object of research is the automized control system of a technological process of the saw timber, woodchip boards and fibreboard guillotine cutting. The purpose is the rise of the use of resources efficiency in the saw timber, woodchip boards and fibreboard guillotine cutting. The rise of efficiency is based on the research and development of system models and cutting maps optimization algorithms which take account of the technological constraints imposed by the equipment. The methods used are the systems analysis, object-oriented simulation, the evolutionary methods of the optimization based on the genetic programming, data parallelism. Scientific results: 1) The mathematical model of the process of the saw timber, woodchip boards and fibreboard guillotine cutting, which unlike the other known themes, takes into account not only cutting material use factor but such technological parameters as: number of measured leftovers, length of the cutting path, number of elementary jobs. 2) For the first time it is developed the GA of the guillotine cutting process optimization. In the developed GA the usage of the new problem-oriented fitness-function, genetic operators and the method of a task coding in a chromosome is proposed. This has allowed to lower material losses about 2-3 % and to decrease the executable manufacturing operations number. 3) The parallel genetic algorithm on the basis of the islands model is developed. The GA proposed allows to reduce material losses up to 5 % at the solving of the high dimensional guillotine cutting problem. Efficient parameters of individuals exchange between subpopulation are defined. 4) The modified parallel genetic algorithm of the optimal guillotine cutting maps search in real time is developed. The GA developed is based on the "worker - master" model. This approach has allowed to use the additional correcting feedback and to keep record of the leftovers in the real time. The practical value is development of the mathematical model of automized guillotine cutting process. It has allowed to reduce a time to preparation of cutting path and note all technology requirements. UML language is applied for the model development; the parallel genetic algorithm developed has allowed to get the optimal solution for the tasks of high dimensionality; the time of guillotine cutting process and the time of the preparatory guillotine cutting process is reduced; the software developed allows to automize guillotine cutting process, to increase the speed of guillotine cutting process, to reduce material and labour expenses of furniture firm. Scientific value of operation consists in: 1)development of the mathematical model of the automized guillotine cutting process taking account of all technology equipment requirements; 2)development of the optimal guillotine cutting process search methods, 3)the improvement of the automized control system of guillotine cutting process by means of additional correcting feedback application. The algorithms, models and techniques, results of researches and the system of optimization of guillotine cutting process and cutting paths developed in the thesis, are used in the furniture manufacturing process by Donetsk regional plant CM "REMMEBEL", by furniture assembly department of Open Company " Synthesis ", and in the educational process and in the research efforts of the "automized control systems " faculty of DonNTU.