Abstract: An improved genetic algorithm that integrates into non-uniform mutation operator and niching technology, which could run toward the optimal solution effectively, and to better maintain the diversity of the solution and high convergence speed was proposed. Meanwhile, taking shovel width and shovel dip for design variables optimizes loading capacity, shovel grubbing force and shovel coaming thickness of roadheader shovel. Optimization results show that: shovel width is reduced by 1.33%, shovel dip is reduced by 9.47%, and Shovel coaming thickness is increased by 11.4%, which has a significant effect on the overall stability of roadheader and the shovel strength. Loading capacity is increased by 4.72%, and shovel grubbing force is increased by 5.12% after optimization, which improve comprehensive performance of the loading mechanism of roadheader and accord with development trend of roadheader shovel in line with the current and future. Meanwhile, the reliability of shovel is analysed under the environment of co-simulation of Pro/E, ADAMS and ANSYS.

Key words: roadheader shovel, loading capacity, shovel grubbing force, shovel coaming thickness, improved genetic algorithm, optimization design, reliability