A survey on metaheuristic approaches and its evaluation for load balancing in cloud computing

Abstract

In daily life there exist many problems whose objective are to either maximize or minimize some value with following some constraints (like load balancing in cloud with aim to maximizing QoS, Travelling salesman problem with aim to minimize total length of trip). These types of problems are optimization problems. Out of these problems there exist many problems which comes under NP-Hard category. To get nearby optimal solution of these problems in polynomial time the metaheuristics approaches are used. Metaheuristics are nature inspired algorithms which provides optimal solution by utilizing combination of exploration and exploitation. This paper provides a survey of Metaheuristic approaches (consisting of need, applications, characteristics, general classification and fourteen approaches under it). Compared all approaches corresponding to key parameters, mechanism. On the basis of literature survey and comparison, cuckoo search has been considered better due to global search via levy flight and generality (because of single parameter setting in cuckoo search). Implemented Randomized algorithm, Genetic Algorithm and Cuckoo Search to solve Load Balancing problem in Cloud Computing with aim to minimize makespan time and proved through results that cuckoo search is better. These experimental results were obtained using CloudSim 3.0.3 toolkit by extending few base classes.