Minimising Numbers of Losses and Abandonments in Small Call Centres Under a Transient Regime

Abstract

We consider the problem of allocating customers to agents in small call centres so that transient performance indicators in terms of expected numbers of losses and abandonments are optimised. To gain insight into the general structure of optimal policies, we start by (i) using backward induction based upon Bellman's equation for a finite-horizon discrete-time model, and (ii) deriving stationary policies for an infinite-horizon continuous-time model with discounting. We then compare the performance of such policies in our original finite-horizon continuous-time model using a first-step analysis applied to the Laplace transforms of the relevant measures.