Environmentally Constrained Naval Search Planning Algorithms Navy SBIR FY2010.1 Sol No.: Navy SBIR FY2010.1 Topic No.: N101-048 Topic Title: Environmentally Constrained Naval Search Planning Algorithms Proposal No.: N101-048-0110 Firm: Adaptive Methods, Inc 5885 Trinity Parkway Suite 230 Centreville, Virginia 20120 Contact: Jim Farrell Phone: (703) 968-8040 Web Site: www.adaptivemethods.com Abstract: The current ASW route planning and asset allocation algorithms suffer from several major shortcomings. First, the current capability provides solutions that are often counterintuitive and have little tactical utility. Second, the operator has virtually no insight or control over the solutions. Third, the current capability emphasizes an overt and offensive ASW posture when developing solutions. Combined, these shortcomings mean that operators are presented with a solution, have no understanding of why it is effective, cannot impact the solution with knowledge and constraints that only the operator knows, and must ultimately make an "all or nothing" decision to follow the generated plan. As a result, the current capability is effectively a black box solution that is effective for a very small set of problems and does not allow the operator to explore and analyze the search space. The existing Navy Mission Planning capability focuses on "overt search" by optimizing a single objective function called the Cumulative Probability of Detection. The proposed solution, Context-Aware Multiple Objective Planning (CMOP), expands its applicability by considering multiple objective optimization where the objective functions are selected based on mission context. Benefits: The key benefits of the Context-Aware Multiple Objective Planning (CMOP) technology are: (1) spatial-temporal and semantic reasoning to dynamically determine optimization objective functions applicable to the current mission context, (2) description length penalized multi-objective optimization to find solutions satisfying multiple objectives while keeping the solutions as simple as possible, and (3) interactive visualization to provide an unprecedented level of insight and control over the optimization process. This approach is useful in any multiple objective optimization where contextual information influences the objectives that should be used during the optimization process. Return