Greedy Strategy and Scheduling Rule Integration for Multi-task Allocation of Coal-gangue Sorting Robots

This paper addresses the issue that the variation in gangue content of raw coal and the complexity of the production process cause nonlinear changes in the arrival rate, position, and particle size of gangue on the sorting belt, which significantly affects the overall performance of coal-gangue sorting robots in practice. By considering the characteristics of the gangue queue and queueing theory scheduling rules, the paper proposes a multi-task allocation method for coal-gangue sorting robots under time-varying raw coal flow. The method uses a greedy strategy and different scheduling rules for multi-arm robots. A multi-task allocation model is established, including a matching matrix, benefit function matrix, and environmental state matrix. The characteristics of the gangue queue and the mechanisms of certain scheduling rules are analyzed to form a set of rule combinations. Finally, using the greedy strategy, the overall benefit of different scheduling rules in different time windows is optimized. Experiments on short (10m) and long (500m) raw coal flow segments show that different scheduling rule combinations have different sensitivities to gangue queue features. The proposed adaptive scheduling method effectively mitigates the impact of time-varying raw coal flow and outperforms existing methods in improving the overall benefit of coal-gangue sorting.