Comparison of binary coded genetic algorithms with different selection strategies for continuous optimization problems

Just like the crossover and mutation operations, selection operation plays an important role in controlling the performances of genetic algorithms (GA). This paper proposes binary coded genetic algorithms (BCGA) with different selection strategies, such as roulette-wheel, exponential, linear transformation, linear ranking selection, binary tournament selection, power-law based probability selection and threshold selection. Furthermore, the effects of these different selection strategies on the performances of the proposed algorithms are compared and discussed by the experimental results on the benchmark instances of continuous optimization problems. The power-law based probability selection and threshold selection are considered as the most possible competitive selection strategies applied in BCGA for continuous optimization problems while binary tournament selection may be the worst strategy.