期刊名称：Eastern-European Journal of Enterprise Technologies
印刷版ISSN：1729-3774
电子版ISSN：1729-4061
出版年度：2016
卷号：2
期号：2
页码：28-35
DOI：10.15587/1729-4061.2016.65456
语种：English
出版社：PC Technology Center
摘要：It is shown that a high percentage of defective rubber-metal articles at the manufacturing process output is caused by neglecting the subsystem parameters connectivity at the design stage separately within the design and technology, as well as between these subsystems.The research is aimed at increasing the rubber-metal articles production stability and improving the rubber products quality through the development and introduction of a new integrated approach to the design and technology parameters optimization.In the general system of the integrated design of rubber-metal shock-absorbers, the subsystems of designs and manufacturing technologies are singled out, and the correlations between the parameters within these subsystems and the parameters of different subsystems are identified.Optimization problems have different objective functions, in which arguments often coincide fully or within certain boundaries. This significantly complicates calculations since the optimization problem in this case is often multiobjective and multiextremal. To solve this problem, the method that involves complex evolutionary optimization by means of a genetic algorithm is applied.For this, new attributes of the genetic algorithm are created. In particular, new star-shaped character models (chromosomes), with internal links between individual parents and flexible constraints on the variation of the latter during optimization are developed. The result is a paradoxical conclusion: there is an additional possibility to perform multi-criteria optimization of the design and manufacturing technology of rubber-metal articles deeper than with Pareto optimization because Pareto optimization involves a single value for all iterations of search of objective functions in the evolutionary optimization, and the arguments on each iteration may differ on some, connection depth-dependant value when using the proposed method.
关键词：rubber-metal articles;parameters connectivity;genetic algorithms;complex character models
