A Flexible Distributed Scheduling Scheme for Dynamic ESG Environments

Abstract : Based on the holonic C2 organizational control architecture (OCA) that models a C2 organization as an integration of multi-level, de-centralized decision making networks, we present a holonic multi-objective evolutionary algorithm (MOEA) that produces robust and flexible distributed schedules within a dynamic ESG mission environment, such as asset break down, appearance of new events, node failures, etc. The lower level units generate multiple local schedules based on local resources, constraints, and interests (objectives). These local schedules correspond to a schedule pool, from which the Operational Unit can assemble a set of ranked L-Neighboring global schedules according to global objectives, and the actual schedule can shift among different stages of alternative schedules in order to adapt to environmental changes. Global feasibility is ensured at the upper level operational unit, while local autonomies are maintained among lower tactical level units due to the characteristics of the proposed holonic organizational control architecture (OCA). The advantage of this scheduling scheme is that it generates multiple neighboring candidate schedules, which avoids the costly replanning process and also minimizes the adaptation cost.