Research On Traffic Unbalance Of The Call Centers

With the development of CTI (Computer Telephony Integration) technology and the further application of IVR (Interactive Voice Response), call centers are enjoying sound development. Compared with the rapid development of application technology for call centers, study on its management seems to have been lagged behind. The rapid development of domestic telecommunication industry has increased its subscribers and popularized its utility, bringing domestic operators great service stress which includes two aspects described as follows:First, the increasing traffic of communication leads to human resource insufficiency and great pressure to operators of telecommunication service.Second, traffic imbalance of communication service brings about the contradiction between low service level as well as decreased operation efficiency in peak hour and low utilization ratio of resource in non-peak hour.Therefore, it is of vital importance for call centers' managers to improve operation efficiency and increase customers' satisfaction at the same time. Scientific analysis for the status of call centers is the prerequisite to solve these problems. Traditional models can't accurately reflect the characteristic of call centers due to their huge variations in functions. Thus new models are in urgent need for systematical analysis.The first and foremost issue for any research is the analysis of customer demand of call centers under discussion. Chapter 2 begins with various theories and methods about traffic prediction. Based on analysis of real data from a large call center, we bring forward a new forecasting theory and set up a traffic forecasting model. Our main idea is to divide the traffic of certain time period of one day into daily traffic and time period ratio. As to daily traffic, we analyze the key effective factors, setup daily traffic forecasting model and revise the forecasting results with an error-handling model. Furthermore, by analyzing historical data statistically, we draw an important conclusion that the ratio has little fluctuation. For all these reasons, the model estimates the time period ratio by the average ration of different days. Finally, the traffic during certain period of one day is estimated by multiply daily traffic and time period ratio. After comparing the results, we find that the result of the traffic forecasting model is better than those of the other similar models.Undeniably, IVR has become an indispensable component of call centers, distributing most part of service traffic and affecting the whole call centers' operation performance. Chapter 3 sets up three indexes to evaluate the IVR operation, which is the agent time saved by IVR, user-path diagrams of IVR and the using complexity of IVR. All these evaluations provide managers with reference to economic benefits, operation details and user satisfaction produced by IVR.There are two methods to improve operation performance of call centers, namely optimize the IVR construction and optimize staff schedule flow.In order to decrease the complexity of the IVR utilization, chapter 4 studies on several methods to optimize IVR menu. Due to current complicated IVR menu system, traditional optimization methods are not available. Probability priority algorithm is introduced to optimize and construct the IVR menu layer by layer. This algorithm is faster and more effective and has been proved to be the global optimal solution. So it is possible to dynamically optimize IVR menu if arrival rates can be accurately predicted.It is widely accepted that human resources could be the main cost of call centers, making multi-skilled CSR (Customer Service Representative) schedule be the most important task of daily affair of call centers. Chapter 5 introduces the method to calculate staff demand which is based on state transfer equations of Markov process and sets up a new staff scheduling model by analyzing data construction, constraints and objectives.