Neural networks for combinatorial optimization problems in logistics

It is not longer sufficient to attack problems in todays complex logistics with conventional methods. Heuristic methods are required in order to achieve suitable results within an acceptable period. Connectionist models seem to balance adequately necessary, possibly conflicting tasks for an optimised system. In this paper the basic idea of Artificial Neural Nets will be outlined with respect to different types of problems in logistics. As an example of practical relevance applications of Artificial Neural Nets for an automatic crane will be explained and the well known Travelling Salesman Problem in the frame of tourplanning for the crane will be discussed.