A New Framework for Job Shop Integrated Scheduling and Vehicle Path Planning Problem

Area of Science:

• Manufacturing Engineering
• Robotics
• Operations Research

Background:

• Traditional manufacturing struggles with flexible demands like small batches.
• Existing Job Shop Scheduling Problems with Transportation (JSP-T) often neglect robot path planning and collision avoidance.
• Flexible manufacturing requires integrated solutions for scheduling, task allocation, and path planning.

Purpose of the Study:

• To develop a unified framework for workshop scheduling, material handling robot task allocation, and conflict-free path planning.
• To minimize the maximum completion time (Makespan) by integrating handling time and robot paths.
• To enhance adaptability to job changes in flexible manufacturing.

Main Methods:

• An extended JSP-T problem model incorporating handling time and robot paths.
• An improved Deep Q-Network (DQN) for dynamic scheduling.
• The Priority Based Search (PBS) algorithm for conflict-free path planning.

Main Results:

• The proposed scheduling algorithm improved Makespan by 9.7% compared to PPO.
• The PBS algorithm successfully generated optimized, conflict-free paths for multiple robots.
• The integrated framework demonstrated improved performance in Makespan and path optimization.

Conclusions:

• The novel framework effectively integrates scheduling, task allocation, and path planning for flexible manufacturing.
• The approach enhances production efficiency and adaptability in dynamic workshop environments.
• This method provides a robust solution for complex robotic scheduling and navigation challenges.