4TH International Congress on Technology - Engineering & SCIENCE - Kuala Lumpur - Malaysia (2017-08-05)

An Elite Pool-Based Big Bang-Big Crunch Metaheuristic for Data Clustering
In this study, we present an investigation of the capability of an enhanced Big Bang-Big Crunch (BB-BC) metaheuristic for data clustering problems. The BB-BC is derived from one of the evolutionary theories of the universe in physics and astronomy. The BB-BC theory involves two main phases (big bang and big crunch). The big bang phase generates a population of random initial solutions whilst the big crunch phase shrinks those solutions to a single elite solution presented by a centre of mass. Our investigation focuses on how effective is the enhanced BB-BC for data clustering. It is enhanced by incorporating an elite pool (of diverse and high-quality solutions) and local search method (e.g. simple descent heuristic), as well as by utilizing implicit recombination, the Euclidean distance, dynamic population size, and elitism. Those strategies provide a balanced search of diverse and good quality population. The investigation is conducted by comparing the proposed elite pool-based BB-BC with the original BB-BC and other similar metaheuristics. The BB-BC is tested on fourteen different clustering data sets. It is found that the elite pool-based BB-BC performs better than the original BB-BC. The incorporated strategies have a greater impact on the BB-BC. Experiments showed that the elite pool-based BB-BC produces high-quality solutions, and outperforms similar metaheuristics (e.g. swarm intelligence and evolutionary algorithms) reported in the literature.
Ghaith Jaradat, Masri Ayob, Ahmad Abu-AlAish, Ibrahim AlMarashdeh, Mutasem AlSmadi,