Article Preview
Top1. Introduction
There are different ways of improving asset management systems and practices. Studies in the domain of transit asset management have shown that asset management outcomes could be improved through the optimal allocation of fleets in transportation networks (Diana et al., 2006) and the timely management of the purchase and retirement of transit vehicles (Khasnabis et al., 2002). Other studies suggest that results could be improved if managers practice timely vehicle replacement (Davenport et al., 2005). Recently, some authors have proposed that asset management could be improved if the allocation of funds among a set of transit agencies were optimized (Ngoa et al., 2018). These optimization solutions all require the use of mathematical models, such as linear programming or non-linear programming, integer, mixed integer and dynamic programming models, to solve mathematical models of the transit vehicles and systems (Ariaratnam et al., 2002; Kozanidis, and Melachrinoudis, 2004).
Moreover, there is the possibility of using Big Data applications to improve asset management at a transit agency. Big Data could be defined as data whose volume, variety, velocity, veracity, and value (Fosso Wamba et al., 2015) make them difficult for a typical software to process. Big data technologies will enable the transit agency to have the capability to better leverage data, because Big data supports the integration, extensibility, and compatibility that exceed anything that traditional approaches offer for railway asset management (Tutcher, 2014).
For example, Swedish railways deployed a pilot of the use of Big data to monitor and manage assets. The solution depends on a centralized cloud storage. All the sources of data within the Swedish railway exported their data to the centralized cloud system. At this location, managers could use advanced analytics software to analyze the data and make decisions. There are potential advantages to this approach. First, a railway company could aggregate heterogeneous sources of information, such as dynamic conditions of the vehicles, geographic information, weather characteristics, and data related to maintenance tasks. Second, huge amounts of data could be captured in real-time, accurately and at affordable cost. Third, the transit system could acquire the capability to make predictions about failure of components and of transit vehicles (Thaduri, Galar, and Kumar, 2015).
However, there are a number of constraints that make it impossible for transit agencies to choose either of these two paths. First, many transit agencies lack the data and the appropriate technologies to perform mathematical programming. Second, many agencies might not even have the manpower to manage and use such mathematical programming tools even if they were available. There are also barriers that mitigate against the deployment of sophisticated technologies such as Big data. These include the enormous investments in time and money to deploy such projects. The huge costs of maintaining thousands of sensors, readers and related technologies, as well as the added costs of hiring new employees to manage such systems would add to a transit agency’s operational costs. Given such hurdles, many agencies prefer to go a different route in improving their asset management efforts.