Binary composite crossover genetic algorithm for locating critical slip surface

Solving slope stability problems requires determining the critical slip surface (CSS) of a slope and its corresponding minimum factor of safety (Min. F), and determining the CSS is a complex optimisation problem. In this paper, we propose a real-coded binary composite crossover genetic algorithm (RGA-BCC) to locate the CSSs of slopes. A reasonable combination of parameters for the binary composite crossover (BCC) operator in the RGA-BCC is determined through six benchmark function experiments. On this basis, the stability of five soil slopes from the literature was analysed using the RGA-BCC in combination with the Morgenstern and Price method, and the results obtained were compared with those in the published literature to demonstrate that the RGA-BCC is able to efficiently determine the CSSs of slopes and their minimum factors of safety. The results of the sensitivity analyses show that having a reasonable setting for the number of control variables is the key to effectively determining the CSS. Moreover, the effects of slope structural characteristics should be considered when setting the number of control variables and population size. Algorithm comparison results show that RGA-BCC has better performance in solving the slope CSS problem compared with differential evolution (DE) and sparrow search algorithm (SSA).