Current-generation AI planning technology lacks effective, flexible techniques for managing the planning process in the presence of changing information, and for coordinating multiple, distributed planning processes. However, these issues have been the focus of theoretical work in AI over roughly the past five years, and similar issues have been studied by researchers developing distributed, real-time operating systems. The goals of our ongoing research project are, first, to adapt relevant techniques from the literature on operating systems and combine them with techniques from distributed and real-time AI in a system that is capable of forming plans in a dynamic, multi-agent environment, and, second, to assess the relative strengths of these various techniques for a variety of environments and tasks. Towards these ends, we are building DIPART - the Distributed Interactive Planner's Assistant for Real-Time Transportation planning - prototype simulation system that includes both a network of cooperating agents and a simulated environment. In this paper, we describe the main objectives of DIPART and the general approach used in building it.