Data driven queueing systems’ capacity estimation with incomplete information.

01 Jul 2023 (modified: 12 Dec 2023)DeepLearningIndaba 2023 Conference Withdrawn SubmissionEveryoneRevisionsBibTeX
Keywords: estimation, capacity, simulation, queueing systems, Krivulin, linear models, overtake
Abstract: Data-driven capacity estimation is one of the most significant issues in systems in which efficiency and effectiveness are key performance indicators. While several queueing models have been put forth in the literature, they typically assume the presence of complete information, i.e., time stamps for arrivals, service starts and departures. In many practical settings, only a subset of these timestamps is available (often consisting of system arrivals and departures). This paper presents simple algorithmic approaches to estimate the capacity of a partially observable queueing system. The observer can only see the arrival and departure times of the customers in and out of the system and wishes to estimate the number of servers (or rate of service). One algorithm is based on Krivulin’s recursion, the other one is based on the number of overtakes and the third is a methodology developed that uses Generalized linear model approach. Using both, synthetic data, and real-life data from the emergency department of a larger urban hospital data, a comparative evaluation of the two algorithms is conducted. Our results show that for a stationary queueing system, our algorithms produce extremely accurate estimates of the number of servers based on rather limited data.
Submission Category: Machine learning algorithms
Submission Number: 2
Loading