A multi-task allocation method for coal gangue sorting robots based on the integration of greedy strategy and scheduling rules

The complex raw coal mining process and the variation in the raw coal gangue content result in nonlinear changes in the arrival time, position coordinates, and particle size of the gangue on the belt conveyor, which affects the overall benefits of coal gangue sorting. Based on a comprehensive consideration of the gangue queue characteristics and queuing theory scheduling rules, a multi-task allocation method for coal gangue sorting robots, integrating a greedy strategy and scheduling rules, was proposed. A basic framework for multi-task allocation of multi-arm coal gangue sorting robots was constructed, including a matching matrix, benefit function matrix, and environmental state matrix. The characteristics of each dimension of the gangue queue information and the mechanisms of some scheduling rules were analyzed. The combination methods of different scheduling rules were studied, and a scheduling rule combination set was established. Through the greedy strategy, the overall benefits of different scheduling rules within different time windows were compared, with the sorting rate and task completion success rate in the coal gangue sorting process serving as the overall benefit rate. The scheduling rule that maximized the overall benefit was selected for multi-task allocation. A time-varying raw coal flow simulation environment with different maximum coal throughput was set up for multi-robot coal gangue sorting task allocation simulation experiments. The results showed that for time-varying raw coal flow samples with maximum coal throughput of 120 and 150 kg/s, the coal gangue sorting rates were 97.69% and 89.10%, respectively, when using the proposed multi-task allocation method with a greedy strategy and scheduling rules, which was an improvement of 6.82% and 5.67%, respectively, compared to the single scheduling rule method. The task completion success rates were 95.64% and 86.46%, respectively, showing improvements of 3.02% and 2.13%, respectively. The standard deviation of robot arm utilization was lower, indicating that the method reduced the impact of the time-varying raw coal flow on the overall benefits of coal gangue sorting.