Grant: Department of National Defence / NSERC Research Partnership

Information monitoring is essential in work-flow execution environments to provide statistical interpretation on executing processes, based on control functions on process parameters. Monitoring aims toward an optimization trade-off between timely response and comprehensive data gathering. The problem usually involves multiple monitoring points reporting to a centralized coordination center. In contrast, distributed monitoring requires physically distributed and autonomous agents bound by local and potentially inaccurate shared information collected in a dynamic and uncertain environment. In this proposal, we aim to elaborate a framework capable of distributed plan execution monitoring to address recent developments on the global arena which require large-scale joint coalition operational plans. These operations need to be adaptive to the dynamic that unfolds during execution as exemplified by search and rescue operations and humanitarian aid distribution. The goal of this initiative includes research contribution objectives in these areas: 1) data exchange protocols; 2) distributed monitoring algorithms; 3) analysis of monitored data; 4) monitoring architecture and design; 5) concept validation through logistics distribution case study. The underlying processes will receive decision vector feedback from distributed monitoring agents, each with its own environment, information domain, goals, resources and capabilities. Moreover, analysis of monitoring data will be used for ongoing adaptive planning efforts using on-line analytical processing (OLAP) and on-line analytical mining (OLAM). These techniques allow identifying emerging patterns or relevant trends. Also, to induce accurate multi-player information exchanges, we propose the use of game theory and mechanism design. The architecture and design will support command and information perspectives necessary to separate different hierarchical levels for localized task execution with respect to space, time, capability and environment. Finally, as test-bed exercise, we will conduct case studies on logistic distribution where the execution of dynamic vehicle routing will be monitored from distributed locations.