Planejamento de equipes de atendimento para o call center CEMIGD baseado em um algoritmo de filas generalizadas

• Ano de defesa: 2016
• Autor(a) principal: Alexandre Henrique Martins
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Call center; Distribuições de probabilidade; Teoria de filas generalizadas; Dimensionamento de equipes; Centros de atendimento ao cliente; Teoria das filas; Engenharia de produção; Distribuição (Probabilidades)
• Link de acesso: http://hdl.handle.net/1843/BUOS-AV3JZJ

The Brazilian power companies are subjected to the federal guidelines established by Agência Nacional de Energia Elétrica - ANEEL. By means of resolutions, this regulatory agency sets the procedures of how the services should be provided focusing on quality and the constant assistance to the clients of those companies. Thus, one of the resilientmanagement issues concerning Cemig D is related to client attendance through call center. This channel is settled to provide assistance through the telephone number 116 and it is subjected to assistance quality parameters in order to answer at least 85% of the calls in up to 30 seconds and to keep the abandoned calls rate at the waiting period up to 4% of the calls. So the attendance teams plaining should be done with the limits set by the national agency on focus. For that reason, a software concerning the scheduling of attendance points was developed using the concepts of General Distribution Queuing System G/G/s in which the parameters of income calls and the attendance rates can be described as a particular probability distribution, also controlling the abandoned calls rate. This software can be compared with the actual plaining pattern of Cemig D Call Center which uses the methodology based on Multi-server Queuing Model M/M/s - and on estimation of parameters by analysis of the number of received calls prediction and by the averageattendance time. The achieved results with this research show that the system of scheduling by General Distribution Queuing reveled to be more efficient, once for each 30 minutes of interval, the distribution of attendees was more balanced than the solution adopted by the company. Furthermore, the attendance quality parameters were not violated and thesolution would generate a larger finance economy on the first semester of 2016 plaining simulation