Hybrid Genetic Algorithm and Linear Programming for Bulldozer Emissions and Fuel-Consumption Management Using Continuously Variable Transmission

This paper develops a hybrid optimization approach combining genetic algorithm (GA) and integer linear programming (ILP) to solve the nonlinear optimization problem of managing the fuel consumption and emissions of a tracked bulldozer. Furthermore, the authors propose that a continuously variable transmission (CVT) can better exploit the efficient zones of the engine maps. The original transmission system of the Caterpillar D6T bulldozer consists of a five-gear transmission, whereas the gear ratios of the proposed CVT are continuous and can be assigned according to transmission design. The fuel consumption and three emission items of the engine, unburned hydrocarbons (HC), carbon monoxide (CO), and nitrogen oxides (NOx), are studied. Vehicle-terrain interactions are formulated and the excavation program is characterized by excavation depth and speed. The target of the multiobjective optimization problem is a combination of fuel rate and three emission items. Results show that, for digging depths less than the bulldozer blade maximum digging depth, the target can be improved by more than 31% using CVT incorporated with GA compared to the conventional transmission, obtained by shifting engine operating points from low efficiency zones to optimum points. Finally, integer linear programming is used in a hybrid manner with GA to solve for the optimum combination of excavation steps in tasks of specified digging depths more than the maximum digging depth of the bulldozer blade. Results show that the proposed method can improve the target value up to 18% with the same digging time, and can improve the target value up to 32% using the hybrid optimization approach without time constraint.