The Problem of Updating Adjustable Parameters of a Grain Harvester in Intelligent Information Systems

The problems of updating adjustable parameters of a grain harvester tools operating in changing environmental conditions is considered. For technological adjustment of such hierarchical multilevel systems, intelligent information systems are applied. While performing technological adjustment of the harvester in the process of harvesting, the incoming quantitative, qualitative, and estimating data are analyzed. When considering semantic spaces of environmental factors and the harvester adjustable parameters, different kinds of uncertainty stipulate application of logical-linguistic approach and mathematical apparatus of fuzzy logics. The paper considers the questions of creating a knowledge base for updating adjustable parameters in cases when a value deviation of the quality harvesting indices from the standard one is observed. Since there are a lot of reasons for a fault, and it is unknown beforehand which of them has resulted in deviation, there are quite a number of ways of responding them. Interrelations between indices of functional efficiency and adjustable parameters are established with the help of empirical rules obtained on the basis of expert data. In order to optimize the fuzzy inference mechanism operation of the intelligent information system there appears the necessity for establishing relevance of the applied rules of knowledge base. To solve this problem a game-theory approach has been used, the concepts of efficiency indices matrix and risk matrix of ineffective decision-making have been introduced. An example of choosing a strategy of searching for adequate response to the occurrence of the harvesting indices fault has been given. Laplace criterion, expected-value criterion, and Savage test used for decision-making in “games with nature” have been applied. The analysis of the obtained results has been carried out, conditions and the sphere of application of the suggested approach have been discussed.