ACS (Air Campaign Scheduler) is a constraint-based air operations scheduler. ACS has been designed for tight integration with external planning systems and feedback from the executing environment, to support continuous, globally motivated control of air operations.

Built on the top of the OZONE  scheduling framework, ACS provides a range of air campaign scheduling capabilities.  In generative mode, it can be used to efficiently generate assignments of aircraft and munitions to a given set of input target/DMPI demands, taking into account target priorities, desired levels of destruction, time-on-target (TOT) windows, temporal sequencing constraints, feasible weaponeering solutions, and aircraft/munitions positioning and availability constraints. ACS can also be used in incremental mode, both (1) to accommodate and integrate new demands into a continuously evolving air campaign schedule, and (2) to reactively reallocate in response to unexpected changes in the execution status (e.g., loss of aircraft, insufficient destruction effect). Finally, ACS provides capabilities for selective (user-driven) relaxation of constraints, providing a basis for exploring alternatives (e.g., delaying missions, surging) in situations where all constraints cannot be satisfied. The ACS scheduler is quite efficient. An assignment for a representative problem consisting of 1,700 DMPIs (involving approximately 5,500 sorties) is solved from scratch on a Pentium II in about 35 seconds; incremental revisions/additions to the schedule are performed in real time. The output from ACS is a set of assigned strike missions designating, for each input target/DMPI demand, the set of sorties to be flown (possibly converging on the target from different bases), the munitions to be expended from each base and precise time windows for various stages of the flight itinerary (including TOT windows).

ACS has been incorporated as a resource allocation and scheduling component of larger-scoped systems for air campaign planning and control in two projects, JFACC and EBO.
 

JFACC PROJECT

Within DARPA’s Joint Force Air Component Commander (JFACC) program ACS has been integrated with SRI’s CPEF planning system to produce JPS (JFACC Planner/Scheduler), a multi-agent system for incremental continuous management of air campaign plans and schedules. ACS is designed to deliver continuous, mixed-initiative allocation and scheduling of air campaign resources to support tight integration with external planning systems and rapid response to unexpected real-world events. These capabilities are intended to support the JFACC program's focus on "the agile and stable control of distributed and dynamic military operations conducted in an uncertain and rapidly changing environment."

Working collaboratively with Dr. Karen Myers and her group at SRI International in the development of an integrated planning and scheduling capability for air operations, we successfully coupled the ACS scheduler with SRI's Air Campaign Planning system (ACP) Within JPS, ACP takes responsibility for translating high-level objectives into resource supportable tasks (e.g., targets to hit, CAP missions) and managing this linkage, and ACS takes responsibility for generating and maintaining resource assignments and task execution times. Experiments with JPS have demonstrated the advantage of early consideration of potential resource allocation problems in eliminating infeasible/unpromising plan options and improving the efficiency of developing an air campaign plan. We have also demonstrated the advantage of more tightly interleaved planning and scheduling strategies in reactive situations (e.g., newly received objectives/targets, changes in available resource capacity), in terms of both improving computational efficiency and maintaining solution stability and minimizing change.

    Publications

Myers, K., S.F. Smith, D. Hildum, P.A. Jarvis and R.D. Lacaze "Integrating Planning and Scheduling through Intensity Adaptation", IJCAI-2001 Workshop Program: Planning with Resources, Seattle, Washington, Aug., 2001

Myers, K., and S.F. Smith, "Issues in Integrating Planning and Scheduling for Enterprise Control", DARPA Research Symposium on Advances in Enterprise Control, San Diego, CA, Nov., 1999

EBO PROJECT

More recently, within the USAF Research Laboratory’s “Effects-Based Operations” (EBO) program, ACS has been integrated with the JTT targeting tool and the CAT campaign assessment tool to provide an initial interactive environment for iteratively developing alternative course of actions, for analyzing various options from the standpoint of resource feasibility and for transforming final results into ATO inputs (mimicking the integration achieved in the initial EBO jumpstart program). Proposed work within the EBO program will focus on deepening ACS’s model to incorporate weight of effort, threat and packaging considerations, and on broadening interaction with CAT and other strategy development/assessment tools.
 

    Researchers
       Qu (Joe) Zhou
       Charles Collins
       Stephen Smith (PI)