Pair Selection of Appropriate Taxi Drivers Using Social Network Analysis Models

Abstract

The current work emphasizes on a Taxi rental company which possess 30 cars. In an effort to benchmark the company’s performance and functionality/usability system, the owner of the company decided to utilize the cars as much as possible in such a way to avoid any Taxi remaining in an idle/inactive status. The company’s system typically was consisted of the following steps as follows: a Taxi car is usually used by a pair of two different drivers within 24 hours so as the first driver takes care of the morning half-day, while the other one takes care of the night half-day. Doing this can help the company to maximum its monetization process leading to optimum revenue and profits. However, one of the problems associated with the current system is that, in case any of the driver pairs will not be able to come to work punctually, then this is going to affect the overall time scheduling of the driving plan for that day leading to time conflict and loss of money for the company. Accordingly, the selection of the appropriate pair of drivers is crucial for the owner of the company. To solve these issues and in order to address the above-mentioned problems, a Process Mining technique based on the Social Network Analysis algorithm was applied and used with the intention of better analyzing and investigating the behavior of the drivers so as to select the best "pair" of drivers for the relevant working days. Subsequently, by using the resulting/generated Social Network graphs/models, the owner of the company was capable of simulating and illustrating the relationships and communicational dependencies amongst the drivers. Due to the fact that the company was using a very traditional way of data collection, therefore, the data was captured and stored manually within a paper-based approach. Nevertheless, this work can provide groundwork for further and future studies and research in such a way that several Process Mining techniques (including Social Network Mining methods) can be applied in versatile scenarios and situations whereas the data is typically captured, gathered and stored manually.