Optimizing multi-machine task allocation for deep loosening operations

Task allocation can dominate the energy consumption in agricultural equipment during operations. Particularly, energy efficiency can promote sustainable development in the context of carbon emissions peaking and carbon neutrality in modern agriculture. Among them, tasks can be programmed into integral parts of operations in multiple agricultural machinery. In the realm of collaborative multitasking, multiple agricultural machines can be utilized to execute the operation tasks within individual fields. The well-defined partitioning of their duties can substantially reduce the overlapping coverage areas. As such, this allocation can then translate into a more effective reduction in energy consumption. Furthermore, partial idle time can be offered for agricultural machinery, when the number of tasks exceeds the available number of machines. The task allocation can also have the potential to enable the simultaneous deployment of multiple agricultural machines within the same field. In this study, a tradeoff was provided for the operational pathway and the number of rows. The operational area was then established for each agricultural machine. The potential conflicts were also effectively resolved in the course of the multi-machine collaboration. Taking the deep tillage operation model as a case study, both effective and energy-efficient collaborative operations were achieved within the agricultural machinery group. To this end, a systematic analysis was performed on the energy consumption during collaborative operations. Subsequently, a multi-objective optimization was proposed to minimize the energy consumption and the completion time of the agricultural machinery group, wherein the task volume was set as the optimization variable. INSGA-II was employed to optimize this model, including the initialization stage, logistic mapping and incorporating, as well as reverse learning. An improved heuristic method of population initialization was employed to enhance the population diversity. In addition, the barrier function was designed to prevent the solution in local optima. Moreover, the uniform crossover operator and simulated single-point mutation were combined to ensure the integrity of integer encoding. The efficacy and practicality of the improved model were systematically evaluated in real-world case studies using deep loosening operations. The results demonstrate that the improved model achieved a notable 4.35% reduction in machinery group energy consumption, and a 4.51% reduction in completion time, compared with the conventional NSGA-II. Furthermore, there was a remarkable 5.51% reduction in the machinery group energy consumption and an impressive 42.65% reduction in the completion time, compared with conventional operations. In essence, the optimization fully met the immediate needs of task allocation across various work scenarios during collaborative multi-machine operations in practical agricultural production. For instance, the allocation of operating rows can be expected to efficiently define the working areas in the seeding and harvesting machines with different working widths, compared with the conventional parallel operation. This innovation can also lead to more optimal solutions in group task allocation.