Real time task planning for order picking in intelligent logistics warehousing.

Area of Science:

• Operations Research
• Artificial Intelligence
• Robotics

Background:

• E-commerce growth necessitates efficient logistics warehousing.
• Intelligent unmanned systems enhance operational efficiency and reduce costs.
• Order assignment and path planning are critical but often separately optimized.

Purpose of the Study:

• To address the joint optimization of order assignment and path planning in real-time task planning (RTTP) for unmanned logistics warehouses.
• To minimize total order processing time and travel costs.
• To develop novel algorithms for enhanced efficiency and effectiveness.

Main Methods:

• Designed a joint optimization evaluation indicator and a joint optimization task planning (JOTP) algorithm.
• Introduced a reinforcement learning (RL)-based approach (JOTP-RL) using Q-Mix for order selection, modeling it as a partially observable Markov decision process (POMDP).
• Optimized the improved THA algorithm for path planning, reducing redundant calculations and considering congestion.

Main Results:

• Proposed algorithms significantly outperform baseline and state-of-the-art methods.
• Demonstrated superior efficiency and effectiveness in execution time and task optimization.
• Validated through extensive experiments on two datasets.

Conclusions:

• Joint optimization of order assignment and path planning is crucial for intelligent unmanned logistics warehousing.
• The developed JOTP and JOTP-RL algorithms provide effective solutions for the RTTP problem.
• The research offers significant advancements in optimizing logistics operations for e-commerce.