Novel Feedback-based Operators in Solving Multi-skill Resource-Constrained Project Scheduling Problem

Multi-skill resource-constrained project scheduling problem investigates a practical situation where resources can perform various tasks. Both the assignment of task-resource and the working sequence on each resource would arise solutions with various total duration and cost. Due to the NP-hard nature of such problem, two feedback-based operators are proposed to improve the quality of solutions. Different from classical operator in evolutionary algorithms, our approaches improve a solution by analyzing the schedule generated and make slight changes on the solution. To be specific, resource-leveling operator is used to balance the workload of resources while schedule-compress operator eliminates idle time on resources. Additionally, an encoding scheme with partial task list and resource list and flexible decoding scheme are introduced to adapt these operators. Experiments have been conducted and result validates the effectiveness the proposed operators.