Akulov N. Modelling of Enterprise Stocks Dynamics in Conditions of Non-stationary Environment

Object - stocks dynamics in productional-distributional system that acts in non-stationary environment. Aim - the development of multi-agent based models of stocks dynamics that are adequate for solving tasks of logistic flows management in non-stationary environment. Methods - generalization and comparison, classification, system analysis, statistical, simulation modelling, econometric. Algorithms and methods of reduction or elimination of wares deficit are offered: multi-agent algorithm of demand forecasting in the case of non-stationarity, reserve stock calculation methods based on pseudo-variation coefficient and based on accounting of forecasting error, the protocol of negotiation between sections of a supply chain for non-deficit supplies that is an upgrade of existing approaches. Scope of analytic method of ABC-analysis is extended, two types of dependencies for groups constraints determination (arctangent and sigma-like curve) are offered. Multi-agent models of supply dynamics are developed; they are software-based complexes for automated simulation experiments and primary results processing and are used for researching of correlations between deficit influencing factors and for decision making in the selection of the best method of non-stationary environment demand forecasting, and in the selection of logistics system deficit minimizing new warehouse location. The series of experiments in research of correlation between total system deficit and quantity of clients in the system, quantity of warehouses, and quantity of delivery vehicles is run over by the developed models. Area - enterprise stocks management, teaching process.