In This paper we discuss about the redundancy allocation problem when uncertainty exists in the problem. The innovation of this paper is the use of fuzzy universal generating function for calculating the availability of system. The system consists of component in series and for each component, elements are used in parallel. The system has the range of levels for performance from complete perfect to complete failure. Identical redundant elements are included in order to achieve a desirable reliability. The elements of each component are characterized by their cost, weight and availability. These elements are chosen from a list that available in market. To find optimum combination of elements for each component we calculate the system availability and then use proposed NSGA-III. The proposed NSGA-III heuristic determines the maximal availability system configuration base on weight constrain.