Information Sharing Impact of Stochastic Diffusion Search on Population-Based Algorithms Mohammad Majid Oudah al-Rifaie – [scale=0.1]0homemDropboxGoldReportLyxThesisLogo.png– Thesis submitted for the degree of Doctor of Philosophy of the University of London Department of Computing, Goldsmiths College January 2012

This work introduces a generalised hybridisation strategy which utilises the information sharing mechanism deployed in Stochastic Di usion Search when applied to a number of population-based algorithms, e ectively merging this nature-inspired algorithm with some population-based algorithms. The results reported herein demonstrate that the hybrid algorithm, exploiting information-sharing within the population, improves the optimisation capability of some well-known optimising algorithms, including Particle Swarm Optimisation, Di erential Evolution algorithm and Genetic Algorithm. This hybridisation strategy adds the information exchange mechanism of Stochastic Di usion Search to any population-based algorithm without having to change the implementation of the algorithm used, making the integration process easy to adopt and evaluate. Additionally, in this work, Stochastic Di usion Search has also been deployed as a global optimisation algorithm, and the optimisation capability of two newly introduced minimised variants of Particle Swarm algorithms is investigated.