Dynamic strategic planning: A hybrid approach based on logarithmic regression, system dynamics, Game Theory and Fuzzy Inference System (Case study Steel Industry)

In this paper we propose a hybrid approach for Dynamic Strategic Planning in the steel industry. We have assumed demand, supply and price as main sub-systems. The interactions among the sub-systems are analyzed through multiple logarithmic regression analysis supported by historical data-base. We use three static and four dynamic scenarios using the leader-follower and coalition game theory paradigm to simulate using system dynamic. The generated simulated data, i.e., initial parameters as inputs, and independent variables as outputs, are used to mine the fuzzy rules of a fuzzy inference system (FIS) to estimate the future behavior of the system. The FIS is used to conduct the most probable cases in the market. The most likely strategy on the basis of long-term behavior of the market, i.e., the average case, is determined. Implementation of the average strategy which comes from the dynamic nature of the parameters, variables, casual loops, system dynamics, game theory, and fuzzy inference system in long-term is reliable. Short-term noises cannot put meaningful impact on the results, as all of them are considered in the procedure of proposed dynamic strategic planning. The whole framework has been applied on a real case study in the steel market.