MAPDP: Cooperative Multi-Agent Reinforcement Learning to Solve Pickup and Delivery Problems

Authors: Zefang Zong, Meng Zheng, Yong Li, Depeng Jin9980-9988

AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct extensive experiments on a randomly generated dataset and a real-world dataset. Experiments result shown that the proposed MAPDP outperforms all other baselines by at least 1.64% in all settings, and shows significant computation speed during solution inference.
Researcher Affiliation Collaboration 1Beijing National Research Center for Information Science and Technology 2Department of Electronic Engineering, Tsinghua University, Beijing, China 3Hitachi (China) Research & Development Corporation
Pseudocode No The paper does not contain a pseudocode block or an explicitly labeled algorithm.
Open Source Code No The paper does not contain any statement about making the source code available or provide a link to a code repository.
Open Datasets No We first generate a random dataset with randomly distributed node locations with demands for efficient performance comparison. ... We collect real-world data from an online logistic platform providing services in Guangdong, China, including more than 100 thousand order pairs within a month.
Dataset Splits No The paper describes using randomly generated and real-world datasets and mentions network training, but does not specify explicit train/validation/test splits (e.g., percentages or sample counts) for reproducibility.
Hardware Specification Yes All experiments are conducted using Pytorch 1.7 on 4 2080Ti GPUs.
Software Dependencies Yes All experiments are conducted using Pytorch 1.7 on 4 2080Ti GPUs.
Experiment Setup Yes The networks are trained via Adam optimizer with L = 3, dk = 128, H = 8 and learning rate lr = 0.001.