Flexible Hybrid Technique for a Fully Fuzzy Mixed Integer Programming: Case Study for a Distribution Supply Chain
Abstract
: Many researchers have studied uncertainty models using fuzzy set tools and techniques. One approach in fuzzy mathematical optimization is full fuzzy linear programming, in which all parameters, including the continuous decision variables, are fuzzy. Many real-world problems involve mathematical models that include integer variables, specifically binary variables, as well as continuous fuzzy variables. Such models, called fuzzy mixed-integer programming models, have not been comprehensively studied so far. In this research, a hybrid method is developed for a fully fuzzy mixed-integer programming model with binary variables. The main part of this model includes the objective function and constraints. The objective function expressed as a trapezoidal fuzzy number, is converted into a crisp multi-objective form. For the constraints, we first convert the left-and right-hand side values into intervals using the well-known fuzzy alpha-cut technique. Then, we convert it into crisp form by defining a value called the degree of constraint satisfaction, which ranges from zero to one. To investigate this problem operationally, we consider a three-level supply chain with factories, warehouses, and retailers, and we implement a mixed model to design the supply chain. The solutions proposed in this research enable the decision-maker to select an optimal solution from a range of values-rather than being limited to a single solution based on the latest information regarding current uncertain conditions. The results from solving the aforementioned problem indicate the hybrid method’s appropriate performance, and the sensitivity analysis on the problem’s key parameters also confirms its reliability.