Sequential Hiring of Contingent Workers Through Learning-Based Optimization

Researchers have proposed a new algorithmic hiring policy, DR-UCB, designed to maximize cumulative profit for firms managing a contingent workforce under conditions of uncertain worker productivity and delayed labor supply, according to a paper submitted to arXiv on June 16, 2026 [1][2]. The policy, formally introduced in a paper titled "Sequential Hiring of Contingent Workers Through Learning-Based Optimization," addresses a sequential workforce management problem where a firm must maintain a team of fixed size while learning which workers are most productive over time [1][2]. The model accounts for two operational frictions: the cost of replacing workers and a random delay before newly hired workers become available, which can stem from prior job commitments, scheduling constraints, or onboarding procedures [2]. The problem is formulated as a stochastic multi-play bandit with costly switching and delayed actions [2]. The DR-UCB policy, short for DelayedReplacement-UCB, operates in learning cycles, using real-time production data to decide when to initiate workforce changes and which workers to replace and hire [2]. The authors demonstrate that the leading-order regret of DR-UCB matches its theoretical lower bound in its dependence on the time horizon, and numerical experiments show it outperforms benchmark policies [2]. The paper was posted on arXiv, an open-access repository for electronic preprints that, as of November 2024, receives about 24,000 submissions per month and has hosted over two million articles since its founding in 1991 [8]. The work appears under the Mathematics > Optimization and Control subject classification and is accessible through the repository's abstract page, which features experimental community tools developed under the arXivLabs framework [1][6]. arXivLabs, launched in 2020, provides a formalized structure for third-party collaborators to build features that enhance the reading and discovery experience, such as citation explorers and code-finding tools, while adhering to values of openness, community, and user data privacy [6][7].