Solving Overlapping Coalition Structure Generation in Task-Based Settings

Authors: Guofu Zhang, Zhaopin Su, Xiaoxiao Song, Zixuan Gao, Miqing Li, Xin Yao

JAIR 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we present the experimental results and address the research questions raised in Section 7.1.
Researcher Affiliation Academia Guofu Zhang EMAIL Zhaopin Su EMAIL (Corresponding Author) Xiaoxiao Song EMAIL Zixuan Gao EMAIL School of Computer Science and Information Engineering, Hefei University of Technology, Hefei, Anhui, China Intelligent Interconnected Systems Laboratory of Anhui Province, Hefei, Anhui, China Anhui Province Key Laboratory of Industry Safety and Emergency Technology, Hefei, Anhui, China Miqing Li EMAIL CERCIA, School of Computer Science, University of Birmingham, Birmingham, UK Xin Yao EMAIL School of Data Science, Lingnan University,Hong Kong SAR, China CERCIA, School of Computer Science, University of Birmingham, Birmingham, UK
Pseudocode Yes Algorithm 1 Checking the feasibility of a disjoint (C, V ).
Open Source Code Yes To ensure verifiability, we provide the source codes of the compared methods and the raw instances at the following link: https://github.com/zgfhfut/OCSG.
Open Datasets Yes Therefore, we generated test instances via simulations of some class of OCSGP. Following the existing work, we generated at total 150 different instances randomly from a normal distribution in the above two settings on the basis of n [10, 100], m [6, 24], r [1, 24], bj k [0, 300], di k [0, 450], and µi [50, 100]. ... To ensure verifiability, we provide the source codes of the compared methods and the raw instances at the following link: https://github.com/zgfhfut/OCSG.
Dataset Splits No The paper describes generating test instances for different scenarios, but does not specify how a larger dataset was split into training, testing, or validation sets. Instead, it mentions generating specific instances and running experiments on them repeatedly.
Hardware Specification Yes All the codes of the compared methods were written in C++ and run on a computer with Intel Xeon E5 2.20 GHz CPU, 32.0 GB of RAM, and Windows Server 2012.
Software Dependencies No The paper mentions that the codes were written in C++, but does not provide specific version numbers for any libraries, compilers, or other software dependencies.
Experiment Setup Yes We adopted the recommended parameter settings in (Liu et al., 2016; Su et al., 2020; Zhang et al., 2020). Specifically, the maximum number of fitness evaluation is 1,500. In GA, the crossover rate is 0.9 and the mutation rate is 0.1. For BPSO, both of the two learning factors are set to 2.0 and the maximum velocity is restricted to 5.0. As for BDE, the crossover probability is 0.25 and the scaling factor is 1.0.