Hybrid Rule-Based Decision Support System to the EOQ Problem in the Form of Posynomial Geometric Programming Formulation with Linear Constraints

The main objective of this paper is to solve theeconomic order quantity problem, which is formulated as a hybrid posynomial geometric programming, using a rule-based decision support system. Avoiding the complexities of the optimization process problems and using the optimum knowledge to build an inference system, which is easier to understand for the decision makers, are the main features of this article. The main approach taken in this paper is to use uncertain decision variables, extracting the optimal knowledge through the hybrid optimization problem and applying this knowledge to design a hybrid inference system. The developed hybrid inference system was applied to 100 random problems and inferred values of the objective function as well as decision variables were compared to the obtained optimum values. Alike decision variables, more than 97% of the deviations between inferred and optimum values for objective function are less than 2%. These results indicated that the developed hybrid inference system is highly efficient to be implemented as an optimized decision support system and its results are quite reliable. Unlike other works in the literature, in this paper, the optimization problem is not replaced with a rule-base which is presented by group of experts. But, an approach is provided to build the optimal rule-based decision support system in which the optimum knowledge is obtained through an optimization problem. This approach will provide decision makers with all optimal decisions that may be needed in the future by replacing the optimal deterministic values for decision variables with the optimal hybrid distribution.