The authors consider the wagon unloading problem which is formalised as a variant of the multimode multiprocessor task scheduling. Two models are designed for the deterministic version of the problem, a mixed integer programming problem and a constraint programming model. Genetic programming is also employed on the design of priority rules, which are applied by a greedy-randomised procedure, the greedy-randomised heuristic is described and how the priority and dispatching rules are used in the genetic programming algorithm. Some computational results are presented with the description of the evolved rules and the performance of the proposed approaches are compared. The performance of dispatching rules and other approaches under uncertainty are presented.