Field Operations Planning for Agricultural Vehicles: A Hierarchical Modeling Framework

The execution of field operations by a fleet of cooperating machines needs to be carefully planned, in order to achieve maximum efficiency. Hence, it is necessary to describe these operations with mathematical models that can be used for optimal planning. The complexity of agricultural operations makes such modeling difficult. Important variables for planning, such as logistics for crop yield, cannot be accurately known in advance. Furthermore, unexpected events may happen during operations, such as harvester blockages and restarting operations. In this paper the overall planning problem is decomposed into a hierarchy of simpler problems that can be solved independently and efficiently. Each lower-level problem is modeled in a way that optimization algorithms from the operations research area can be adopted for their solution. The dynamic nature of the problem and the inherent uncertainties of many parameters suggest the adoption of a closed loop control system, which results in a sequence of planning, execution and re-planning. The resulting optimal operation plans can be complex and their execution requires an advanced degree of machine, or field robot autonomy. Overall, the adoption and modification of existing methods from other research areas such as logistics, routing, factory scheduling, and robotics offer a very promising approach for the efficient planning of operations executed by agricultural vehicles.