A fuzzy-based approach to evaluate multi-objective optimization for resource allocation in cloud

Effective resource allocation can be used to achieve two important parameters in Cloud viz. energy efficiency and data center performance. Multi-objective optimization is one of the techniques to address the issue of resource allocation with the multiple objectives. In this research, we aim to address the issue of resource allocation through a weighted sum based multi-objective optimization technique. In weighted sum method, coefficient is attached with each objective as a user's preferences to decide a priority of objective. Genetic algorithm and fuzzy logic are the identified methods to calculate the co-efficient to generate Pareto optimal solutions. In this paper, we use fuzzy logic to generate the random value of objectives' co-efficient. The proposed fuzzy-based computing is implemented and experimental results show the proposed scheme efficiently generates a random coefficient that assigns priority by considering characteristics of host. Results depict the average improvement in performance by 25.7% in power and 3.67% service level agreement (SLA) violations over the period of 24 hours. Further, it demonstrates that the weight generated gives Pareto optimum solution that points to strict Pareto curve.