An Optimization-based Multi-level Asset Allocation Model for Collaborative Planning

Abstract : Motivated by the Navy's emphasis on networked planning capabilities in maritime operations centers (MOC), we have developed an agent-based multi-level resource allocation model that takes high level commands from the human planners and then dynamically allocates the lower-level assets and processes tasks to accomplish the mission objectives. The agent-based model supports a controllable, multi-player, real time collaborative planning environment in a Windows environment. The architecture allows for adding constraints on and manipulations of organizational structures, such as authority, information, communication, resource ownership, task assignment, as well as mission and environmental structures. The planning problem is formulated as a multi-level optimization problem of minimizing the overall difference between the human specified performance measures and expected performance measures which are evaluated based on how well the assigned resources match the required resources, subject to a number of real-world planning constraints on assets. We applied a Dynamic List Planning algorithm (DLP) to solve the intractable multi-level resource allocation problem. The near-optimal DLP method can generate high-quality solutions in seconds compared to days taken by the branch-and-bound-based search methods.