Algorithmization of Interaction of Components of Expert Virtual Resource of Procedural Type in Managerial Decision-Making Optimization

Abstract

For intellectualization of managerial decision making support in corporate systems it is recommended to use expert and virtual resource (EVR), which is a combination of expert and computer resources. In the case when knowledge corporate environment is poorly structured and it has obscure connections and a multi-level subordination hierarchy, the EVR of procedural type intended for intellectual support of decision-making, is used. Considering the existing diversity of EVR of procedural type, there is a necessity to develop algorithms for interaction of their components in managerial decision making in various modes. Algorithmization of a virtual mode of interaction is related to the sequential use of all the three types of virtual experts, thus the basic procedure of multialternative optimization used by multialternative virtual expert (MVE) prevails. The principal task of developing the algorithm of dual mode interaction between the virtual and real experts in the environment of managerial decision making consists in development of a matrix game. Algorithmization of a collective mode of interaction of real and virtual experts is based on a man-machine procedure which provides a dialogue with real experts on the basis of automatically offered questions, with further formalization of expert answers. An iterative principle of building of this procedure by means of immersion of question-answer process with a team of real experts in randomized environment, is proposed. A Team mode with a dominating expert is based on agreeing the assessments of priority of alternative optional managerial decisions by that expert with assessments by a team of peer experts. The developed algorithms can be integrated into the single environment of the managerial decision making subsystem and used as a means of intellectual support.