Technology Development in the ARPA/RL Planning Initiative

Since its inception, ARPI has supported a number of technology development efforts focused on crisis action planning. In this article we provide a synopsis of the progress made in a selected subset of the technologies that will need to contribute to making the ARPI vision a reality. We focus on ARPl-supported research that is providing critical advances in the enabling technology areas of plan generation, scheduling and temporal reasoning, plan analysis, and overall supporting infrastructure. We conclude with a discussion of several important remaining technology challenges. Motivation and Overview One of the driving requirements for ARPI is that research and technology development must demonstrably address a critical aspect of ARPI vision. The technology development needs fall into 4 key areas: scheduling, plan generation, simulation/projection, and uncertainty management. These are accompanied by supporting development needs in infrastructure. The roadmap also attempts to capture the rate at which emerging technology might reasonably be expected to be used in an Integrated Feasibility Demonstration (IFD). In such demonstrations, prototypes embodying the technology are expected to be applied to a critical problem in a relevant domain, usually in cooperation with other prototypes developed as part of the Initiative. The technology demonstrated in the 1992 IFD focused on predictive aspects of the planning process, i.e., the ability to support a priori development of a military transportation plan by an individual at a central site, based on a single assessment of the situation. The technology capabilities were expected subsequently to be broadened enough to support a demonstration involving a collaborative planning effort to respond to an evolving crisis scenario; this necessitates tight integration of planning, scheduling, and decision support capabilities in an environment that facilitates the collaborative planning process and visualization of partial plans. In this and later phases, technology for distributed collaborative planning will play an increasingly important role, as will the need for more robust techniques for planning and scheduling under uncertain and incompletely specified conditions. In the next sections we present synopses of ARPIsponsored research in the areas of plan generation, temporal reasoning and scheduling, plan analysis, and supporting information infrastructure.