This thesis presents and evaluates the DACS methodology for designing agent-based systems for (non-real-time) production control. This methodology is appropriate and sufficiently prescriptive for a control engineer with only minimal training in agent concepts and with no prior experience in agent development to design an agent-based production control system. This was achieved by deriving the concepts of the methodology from an analysis of the application domain and by specifying all relevant agent-oriented design rules in terms of these concepts. In particular, the methodology consists of a method for analysing the decision making necessary to control a production system; a method for identifying appropriate production control agents, which also includes a rule specifying when to abandon an agent-oriented design approach; and a method for selecting interaction protocols to resolve any decision dependencies between the agents which allows to re-use existing interaction protocols. To evaluate the methodology, several case studies and reviews were performed. In particular, two case studies with students applying the methodology to a realistic test case and two reviews by engineers, who design control systems, were conducted. The feedback was unequivocally that the methodology is appropriate and sufficiently prescriptive for designing agent-based production control systems and that it could be applied in industrial projects in order to gain more experience with the methodology.