An Evaluation of the NSGA-II and MOCell Genetic Algorithms for Self-Management Planning in a Pervasive Service Middleware

Planning (for example choosing most suitable services for self-configuration) is one important task in self-management for pervasive service computing, and can be reduced to the problem of multi-objective services selection with constraints. Genetic algorithms (GAs) are effective in solving such multi-objective optimization problems, and are one of the most successful computational intelligence approaches currently available. GAs are beginning to be used in planning for self-management, but there is a lack of comprehensive work that evaluates GAs performance and solution quality, and guides the setting of GAspsila parameters.This situation makes the application of GAs difficult in the pervasive service computing domain in which performance may be critical and the settings of parameters may have big consequences for performance. In this paper, we will present our evaluations of two GAs, namely NSGA-II and MOCell, in the GA framework JMetal2.1, for achieving multi-objective selection of available services. From these evaluations, suggestions on how and when to use NSGA-II and MOCell are given in the planning for self-management.Our experiences show that to get a true Pareto front for a problem, combining solutions set from different GAs is abetter way than using a single GA.