Determination of optimal variable-sized multiple-block appointment systems

The single-block appointment system is the most common method of scheduling ambulatory care clinics today. Several studies have examined various appointment systems ranging from single-block appointments on one extreme to individual appointments on the other, and including mixtures of these such as multiple-block (m-at-a-time) and block/individual systems. In this paper we analyze a general single-server multiple-block system, one permitting blocks of variable size. In the analysis we use a dynamic programming approach, with some modifications to compensate for the non-Markov nature of the problem. Analytical results and approximations which significantly reduce the computational requirements for a solution are obtained. Examples demonstrate that under certain weightings of the criteria of waiting, idle, and overtime, the generality of the system considered here allows performance superior to that of other commonly used systems.