Optimal Complex Task Assignment in Service Crowdsourcing

Authors: Feilong Tang

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

Reproducibility Variable Result LLM Response
Research Type Experimental Comprehensive experimental results demonstrate that our WMST model and AE-TA algorithm significantly outperform related proposals.
Researcher Affiliation Academia Feilong Tang Department of Computer Science and Engineering, Shanghai Jiao Tong University, China tang feilong@sjtu.edu.cn
Pseudocode Yes Algorithm 1: AE-TA Algorithm
Open Source Code No The paper does not provide an explicit statement or link for the open-source code of the described methodology.
Open Datasets No The paper uses simulated data and internally collected real data (e.g., "We simulated 1500 workers", "We chose 70 articles"), but does not provide concrete access information (link, DOI, citation) to a publicly available or open dataset.
Dataset Splits No The paper does not explicitly provide training/test/validation dataset splits (e.g., percentages, sample counts, or specific split names) needed to reproduce the experiment.
Hardware Specification No The paper does not explicitly describe the hardware (e.g., specific GPU/CPU models, memory) used to run its experiments.
Software Dependencies No The paper does not provide specific software dependencies (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup Yes The timeslot was set as one hour. We simulated 1500 workers, and built one WMST for each worker. In each WMST, weights of skills in leaf nodes were randomly set in [0, 1] and weights of non-leaf nodes were computed using formula (7). All WMSTs had the same structure, with a maximum depth of 5, and each node except the leaf nodes had five children. Tasks were generated at an average speed of µ=10 tasks /hour. The required number gt of workers for each task t was randomly initialized such that gt [5, 50]. The deadline dt was set by adding bt with a time drawn from a normal distribution with a mean of 50 and a variance of 20. Any task randomly requires 1 to 5 skills in leaf nodes of WMSTs of the workers who involved in this task. Let task execution time follow the power law distributions p(x)=cx α, where c and α were randomly selected in c [10, 30] and α [1.5, 2.5]. We set ρtj=0.8 for all tasks.