HS-CAI: A Hybrid DCOP Algorithm via Combining Search with Context-Based Inference

Authors: Dingding Chen, Yanchen Deng, Ziyu Chen, Wenxing Zhang, Zhongshi He7087-7094

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, (iii) we prove the correctness of our algorithm and the experimental results demonstrate its superiority over the state-of-the-art. Empirical Evaluation In this section, we first investigate the effect of the parameter t in the context evaluation mechanism on HS-CAI. Then, we present the experimental comparisons of HS-CAI with stateof-the-art complete DCOP algorithms.
Researcher Affiliation Academia Dingding Chen,1 Yanchen Deng,2 Ziyu Chen,1, Wenxing Zhang,1 Zhongshi He1 1College of Computer Science, Chongqing University, Chongqing, China 2School of Computer Science and Engineering, Nanyang Technological University, Singapore {dingding, chenziyu, zshe}@cqu.edu.cn, ycdeng@ntu.edu.sg, wenxinzhang18@163.com
Pseudocode Yes Algorithm 1: Hybrid phase for agent ai
Open Source Code Yes All evaluated algorithms are implemented in DCOPSovler1, the DCOP simulator developed by ourselves. 1https://github.com/czy920/DCOPSovler
Open Datasets No For each experiment, we generate 50 random instances and report the average of over all instances.
Dataset Splits No The paper describes generating "50 random instances" but does not specify how these instances were split into training, validation, or test sets.
Hardware Specification No The paper does not specify any hardware details used for the experiments.
Software Dependencies No All evaluated algorithms are implemented in DCOPSovler1, the DCOP simulator developed by ourselves.
Experiment Setup Yes Therefore, we set t = (dmax)ρh. Moreover, we choose k = 6 and k = 10 as the low and high memory budget for MB-DPOP, HS-AI, HS-CAI(-M) and HS-CAI. Thus, we choose ρ to 0.25 for HS-CAI(k = 6) and 0.45 for HS-CAI(k = 10) in the following comparison experiments.