We consider a manufacturing job shop with automated guided vehicles. Such a system is characterized by several complex features such as the random pattern of material flows between machines, the limited capacity of machine buffers, shop deadlock phenomena, floating bottlenecks, etc. We will describe a heuristic procedure for the control of materials flows and automated guided vehicles in such a system. The procedure is based on the idea of workload balancing; it tries to balance the workload among machines as well as the workload between the machines and the transporters. A deadlock prevention algorithm is embedded in the procedure. We then present the results of simulation experiments to evaluate the performance of the proposed procedure.