Analiza in napovedovanje zasedenosti klicnega centra

Briški, Lea (author), Demšar, Janez (mentor)

Namen tega dela je preučiti, modelirati in optimizirati delovanje klicnega centra telekomunikacijskega podjetja. Uspešnost delovanja klicnega centra je odvisna od dolžin čakalnih časov, ki se podaljšajo, ko je sistem preobremenjen. Obremenjenost klicnega centra je odvisna od količine dohodnih klicev, zato smo zgradili napovedne modele za napovedovanje števila dohodnih klicev in jih testirali. Klicni center smo analitično obravnavali kot strežni sistem. Izračunali smo kazalnike delovanja po M/M/c modelu in preverili, ali ustrezno modelira delovanje klicnega centra na podlagi primerjave izračunanih vrednosti z realnimi vrednostmi iz podatkov. Ker način delovanja klicnega centra krši nekatere predpostavke M/M/c modela, smo delovanje klicnega centra simulirali z diskretnim dogodkovnim simulatorjem, ki omogoča bolj fleksibilno simulacijo sistema.

• Language: Slovenian
• Keywords: klicni center, regresija, posplošeni linearni modeli, strežni sistemi, M/M/c, diskretni dogodkovni simulator
• Work type: Bachelor thesis/paper
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2024
• PID: 20.500.12556/RUL-161733
• Publication date in RUL: 13.09.2024

Secondary language

• Language: English
• Title: Analysis and Prediction of the Call Center Load

Abstract

The purpose of this work is to analyse, model, and optimize the call center of a telecommunications company. One of the performance measurements of the call center effectiveness is waiting times, which increase in length when system is overloaded. The workload of the call center depends on the amount of incoming calls. Therefore, we built and tested prediction models for predicting the number of incoming calls. Call center was analytically modeled as a queueing system. Performance measurements were calculated based on M/M/c model and compared to real data to see if M/M/c model models call center operations well. Due to some violations of the assumptions of the M/M/c model, we decided to simulate the call center's operations using discrete-event simulation.

• Keywords: call center, regression, generalized linear models, queueing systems, discrete-event simulation