An Extension of a Dynamic Heuristic Solution for Solving a Multi-Objective Optimization Problem in the Defense Industry

Abstract

Project scheduling in a real-life scenario often involves multiple-criteria decision-making in which no single solution exists. To solve such a problem, a multi-objective optimization method has been applied to define the satisfying trade-off between different criteria. In this paper, we focus on a specific use case in the defense industry in which the overall mission is to generate a maintenance plan for the transfer operations of power grid consumers to the new service area. The project objectives include restricting the outage duration during transfer operations, grouping operations concerning the proximity between them, moderating the allocation of supporting resource, and regulating human resources intervention outside business hours. To solve this problem, we propose a combination of heuristic approaches starting by defining a sequence of activities based on their complexities to be scheduled. Concerning the obtained order, a serial-schedule generation scheme (S-SGS) is then implemented by iterating through each activity to define the best time period to proceed the operation in accordance with project’s multiple objectives. Finally, the output is transferred to our existing parallel scheme-based solver, Optimizio, to finally justify the project planning. The proposed S-SGS solution provides a feasible schedule of 110 transfer operations in 2 s with solution evaluation analysis and information of a Pareto frontier in approximately 15 min. The set of Pareto optimal solutions allows the expert to explore potential trade-offs between criteria. Together with a fast execution time of the algorithms that benefits a multi-scenario simulation, our tool demonstrates a potential capacity to get the optimum outcome of the multi-objective optimization project.