Problem of electronic components classifying

Forming the on-board equipment of the space systems with fail-tolerant electronic devices is one of the most important problems of the modern space industry. The first issue is to supply the space industry companies with fake and low-quality production which does not correspond to the reliability requirements. In this paper, authors consider the problem of improving quality of national production of electronic devices. In production of electronic units, using the electronic devices with equal values of characteristics is preferable. The equality of the characteristics is reached if a production lot of the electronic devices consists of devices produced as a single production batch. In the case when the production conditions are unknown, the only way of checking if the whole lot belongs to a single production batch and improving quality of electronic unit and the whole system is complex testing of the whole lot. This problem is the most actual and important for construction of electronic units of the space systems. The spacecraft contains approximately 100000-200000 electronic devices such as electronic chips, transistors, diodes, capacitors, relays, resistors etc. On-board equipment in the space cannot be repaired. Thus, the reliability of such equipment must be maximal. Required reliability level can be reached via many various factors. One of the most important factors is to use the hireliability electronic devices. In this article, authors propose a problem statement of learning the quantity of the production batches in a lot of the devices. This problem is solved as series of cluster analysis problems by special genetic algorithm.