Teoria das filas e delivery online: proposta de um modelo generalizado para sistemas com entradas não homogêneas

Abstract

Queueing Theory is a powerful tool in aiding the management of service quality, however many real problems have different particularities which algebraic solutions do not fully meet, such as the case of non-homogeneous arrivals queues. This study proposes a combination of two methods: one that estimates the intensity of arrivals in the system when it follows a nonhomogeneous Poisson Process and another that allows calculating the probabilities of queue systems with time dependent arrival rates, multiple servers and a limited number of simultaneous users. The methodology is developed based on the study of the challenges faced by an online delivery platform, which mediates services of a variety of businesses and therefore we obtain a generalized model, enabling analysis of various systems.