Introduction Inverse Cycloid Cross Section for Optimal Design of Channels by using Artificial Intelligence

The shape of the artificial channel sections and their optimal design by artificial intelligence methods can have significant effects on lowering the costs of constructions. In this study, an artificial channel with inverse cycloid and its optimal design by bat algorithm was introduced for the first time. Two schemes of sections were proposed: inverse cycloid without horizontal bed and inverse cycloid with horizontal bed. For optimal designing of proposed section, two scenarios were defined: channel with fixed freeboard and channel with variable freeboard. In both scenarios, Manning equation was used as a constraint and Horton equation was used for considering equivalent roughness. Based on sensitivity analysis, optimal parameters of the proposed algorithm were determined. Bat algorithm has been implemented randomly for each section and for each scenario for fifteen times, and coefficient of variation and speed of convergence were estimated. To find out the accuracy of bat algorithm in the global optimum determination, a comparison between solutions of bat algorithm and lingo software was made. Eventually, results from optimal design of proposed sections were compared with trapezoidal channel, parabolic with horizontal bed channel, ellipsoid channel, and general ellipsoid channel. Results from random runs of bat algorithm indicated that coefficient of variation for defined sections and scenarios were about 0.0002 to 0.0013. Convergence curves showed that the algorithm was convergence for all sections and scenarios in 1000 repetition. Findings of bat algorithm and lingo software also demonstrated high accuracy of mentioned algorithm in determination of global optimum. Furthermore, using proposed sections, compared with other common sections, led to 34.7 percent decrease in construction costs. Among sections and scenarios investigated here, the inverse cycloid under second scenario was found more economical, up to 17.15 percent.