Improving Genetic Algorithms for Optimal Land-Use Allocation

Abstract

Land-use allocation (LUA) is a spatial optimization problem for urban planning in the future. The solution to this problem could be enhancing the effectiveness of optimization algorithms that create a balance between urban needs and efficient LUAs. This study will improve the performance of conventional genetic algorithms (cGAs) to address future LUAs. The enhancements will include improved initialization and mutation operators and a modification in the fitness function by the introduction of the area size deviation (ASD). The study will consider four objective functions to be maximized that include suitability, compatibility, conversion cost, and spatial compactness. The improved algorithm will be tested and applied to Taiz, Taiz Governorate, Yemen, to acquire the optimal LUA based on their 2035 plan. The findings produced three scenarios that were objectively weighted. The experimental results revealed that the improved genetic algorithms (GAs) were superior to the cGA for convergence speed, solution efficiency, optimality, and fulfillment of all constraints that included the LUA area size. The study exhibited the impact of weight variations on optimum LUAs for decision makers. This optimization resulted in a significant radial plan that featured green wedges and a smart arrangement of land-uses (LUs).