Rational Queueing
Abstract
This book provides a comprehensive treatment of queueing systems in which customers make strategic decisions about whether and when to join. It covers observable and unobservable queues, priority mechanisms, and information effects, establishing that arrival rates are determined by customer utility calculations rather than being externally fixed.
★ Curator Summary
Hassin's work demolishes a core assumption of traditional WFM planning: that arrival rates are exogenous (fixed by external factors). In reality, customers are strategic agents who decide whether to contact based on expected utility: U = Value - Cost(Wait). When AI reduces wait times or effort, more customers find it worthwhile to contact — the arrival rate itself increases. This is the queueing theory formalization of why Erlang-C systematically understaffs after AI deployment.
Why It Matters
Traditional Erlang-C assumes lambda (arrival rate) is fixed. Rational queueing proves it's not — it's a function of system state. When you deploy AI and reduce wait times, the equilibrium arrival rate increases. This means your post-AI staffing model needs to use a higher lambda than your pre-AI baseline, not a lower one. The math is straightforward: as service rate mu increases, the utility threshold drops, and customers with lower-value inquiries start contacting.
Caveats
The models assume fully rational customers with perfect information about queue state, which is a simplification. Real customer behavior involves bounded rationality, habit, and information asymmetry. The book is mathematically rigorous but the applications to contact centers specifically require domain translation.
Discussion
Loading...