Multi-agent systems (MAS) offer new perspectives compared to conventional, centrally organised architectures in the scope of supply chain management (SCM). Their structure inherently meets the requirements of decentralised supply chains, whereas conventional SCM systems are often restricted in terms of dynamic behaviour, handling severe disturbances at supplier sites as well as dealing with highly customised or complex products. Since necessary data are not available within the whole supply chain, an integrated approach for production planning and control taking into account all the partners involved is not feasible. In this paper a MAS architecture integrating various intelligent agent systems is presented to address the problem.