Brain-Inspired Experience Reinforcement Model for Bin Packing in Varying Environments

Area of Science:

• Artificial Intelligence
• Operations Research
• Computational Neuroscience

Background:

• The bin-packing problem (BPP) is an NP-hard combinatorial optimization challenge.
• Existing methods struggle with BPPs involving random item numbers and shapes in diverse scenarios.
• Maximizing bin utilization is crucial for efficiency in BPP.

Purpose of the Study:

• To develop a unified, adaptive model for optimal bin-packing decision-making.
• To improve bin utilization by mimicking human experience-based reasoning.
• To address BPPs in dynamic and unpredictable environments.

Main Methods:

• A novel brain-inspired experience reinforcement model is proposed.
• The model integrates knowledge representation for information processing and experience storage.
• Knowledge extraction modules train reasoning strategies and enhance decision performance.

Main Results:

• The model demonstrated adaptability to varying environments and item characteristics.
• Performance was evaluated on BPP instances with random item numbers and shapes.
• The proposed model achieved superior results compared to state-of-the-art methods.

Conclusions:

• The brain-inspired model offers an effective solution for complex bin-packing problems.
• The approach provides a unified and adaptive strategy for optimizing bin utilization.
• This research opens new avenues for applying neuro-inspired computing to optimization challenges.