Towards Evolutionary Multitasking
Abstract
The booming popularity of cloud computing (CC) services presents the challenge of effectively dealing with multiple jobs that may be received from multiple customers at the same time. In such a setting, imagine a cloud-based on-demand service that provides customers with access to state-of-the-art optimization tools. Albeit unknown to the service provider, there may exists underlying synergies across various problems that could potentially be harnessed to accelerate the optimization process. However, it is noted that the traditional methods for optimization, including the population-based search algorithms of Evolutionary Computation (EC), have generally been focused on efficiently solving only a single optimization task at a time. It is only very recently that a new paradigm in EC, namely, Multifactorial Optimization (MFO), has been developed to explore the potential for evolutionary multitasking. MFO is found to leverage the scope for implicit genetic transfer across problems in a simple and elegant manner, thereby, opening doors to a plethora of new research opportunities in EC, dealing, in particular, with the exploitation of underlying synergies between seemingly distinct tasks. In this talk, a formalization of the concepts of MFO shall be introduced. Thereafter, the efficacy of the associated algorithm shall be substantiated via a variety of computational experiments in intra and inter-domain evolutionary multitasking.
Authors: Yew-Soon Ong
Published in: International Symposium on Information and Communication Technology (2015)