Best of both worlds: Mitigating imbalance of crowd worker strategic choices without a budget

摘要

Crowdsourcing has become a popular paradigm for requesters to hire ubiquitous crowd workers. The worker’s selfish instinct of choosing the most profitable task can cause the imbalance of task completion: some tasks achieve a number of redundant worker choices, while others may receive no worker response. Although budget-based incentives can mitigate the imbalance of crowd workers’ strategic choices, the extra budget makes them less attractive. To mitigate task completion imbalance without a budget, a price mediation mechanism is proposed. This mechanism works by allowing the crowdsourcing platforms to implicitly adjust task prices, thereby eliciting workers to balance their choices. The price adjustment should be carefully designed to satisfy (1) task completion integrity and (2) no extra budget, while it maximizes social welfare. We prove that this optimization problem is NP-hard to solve. By designing bound function and pruning strategies, we propose an optimal branch-and-bound algorithm for small-scale instances. To further improve the scalability for large-scale instances, a heuristic method based on price transfers is proposed. Experimental results on a real dataset show that compared with benchmarks, our approaches are effective for maximizing social welfare and are beneficial to both requesters and workers.