An Efficient Genetic Algorithm Paradigm for Discrete Optimisation of Pipeline Networks

Engineering and science disciplines make use of genetic algorithms. A number of genetic-based search paradigms have been developed and applied to different problems. A growing demand for algorithms to solve new problems and a never-ending process of designing algorithms strongly suggests the need for more efficient and more robust geneticbased optimisation techniques. Ideally these algorithms should make few assumptions regarding the objective functions and use as little domain knowledge as possible. In this paper, different genetic-based search paradigms including the standard GA, the messy GA and the fast messy GA are compared for the discrete optimisation of pipeline networks.