THUẬT TOÁN TIẾN HÓA GIẢI BÀI TOÁN LẬP LỊCH DỰ ÁN VỚI TÀI NGUYÊN GIỚI HẠN (MS-RCPSP) VÀ ỨNG DỤNG TRONG VIỆC LẬP KẾ HOẠCH SẢN XUẤT THÔNG MINH

Abstract

This paper introduces an intelligent production planning approach that addresses the challenges of scheduling project execution with limited resources in the MS-RCPSP (Multi Skill-Resource Constrained Project Scheduling Problem). Due to its classification as an NP-Difficult problem, finding an exact solution within a polynomial time frame is impossible. However, the paper proposes a method called MEM-RES, which combines the Memetic algorithm with the Reassignment method to find approximate solutions efficiently. The algorithm iteratively improves the solutions with each evolutionary generation. The proposed algorithm has experimented using actual production data from the TNG industrial sewing line. The results demonstrate the algorithm's suitability for the problem and its ability to achieve high efficiency. The solution obtained through this approach provides a schedule that organizes resources for executing each project's tasks over time until finished. Consequently, this solution can be made the automatic production planning, replacing the traditional methods currently employed in production factories, especially those utilizing a production line model.