Abstract

Allocating tasks to the best-fit candidates is a classical problem in crowdsourcing (CS). Most of the existing approaches assume that the task and candidate knowledge is known in advance and ignore the effect of enrolled candidates' willingness on the CS system's selection decision. For instance, an unwilling candidate assigned to a task may quit without completing it, thus depreciating the utility of the CS platform. In practice, a task or candidate may arrive or leave the CS system dynamically. Moreover, a complex task may be broken into smaller sub-tasks, each requiring a variety of computations and expertise. To overcome these challenges, based on a greedy algorithm, we propose a novel approach for skill-oriented dynamic task allocation with willingness factor for complex assignments (SWill-TAC). This approach iteratively attempts to delegate candidates (workers) to tasks depending on the skills required for executing the tasks and the candidates' skill set. SWill-TAC also considers the willingness of eligible candidates and keeps track of the budget constraints of tasks. Finally, the feasibility and efficiency of our approach are demonstrated using the UpWork dataset. Experimental results show that SWill-TAC outperforms Online Greedy, TM-Uniform, Random selection-based, and Minimum payment-based task allocations in terms of the completed tasks count, the utility gained, and success ratio.

Department(s)

Computer Science

Comments

National Science Foundation, Grant 1725755

Keywords and Phrases

Crowdsourcing; Dynamic Task Allocation; Greedy Approach; Skill-oriented; Willingness

International Standard Serial Number (ISSN)

2576-6813; 2334-0983

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2024 Institute of Electrical and Electronics Engineers, All rights reserved.

Publication Date

01 Jan 2021

Download

Full Text Link

Share

 
COinS