Reducing the search space in genetic algorithm : An application in emergence of cities

Abstract

Scientists in the field of spatial economics have proposed different theories on how cities are emerged. These theories were transformed into different algorithms on emergence of cities. Assessing the efficiency of the final emergence from these algorithms is best performed where an ideal or optimal emergence is available for comparison. However, without performing exhaustive search, determination of optimal emergences from an arbitrary setup is almost impossible. This work is an application of genetic algorithm in determining an optimal emergence from a given setup. Ten random initial setups were generated based on power-law distribution and Zipf's law. They were used in simulation of a particular emergence algorithm. Genetic algorithm was then applied to determine optimal emergences from these setups. The very large search space in this problem prevented genetic algorithm from finding suitable emergences. The solution is found by using local knowledge of individuals to reduce their search space. The work affirms the benefit of genetic algorithm in emergence of cities and introduces a method to reduce search space in this particular area.