Chi-Kang Lee

Neural network based temporal feature models for short-term railway passenger demand forecasting

Accurate forecasts are the base for correct decisions in revenue management. This paper addresses two novel neural network structures for short-term railway passenger demand forecasting. An idea to render information at suitable places rather than mixing all available information at the beginning in neural network operations is proposed. The first proposed network structure is multiple temporal units neural network (MTUNN), which deals with distinctive input information via designated connections in the network. The second proposed network structure is parallel ensemble neural network (PENN), which deals with different input information in several individual models. The outputs of the individual models are then integrated to obtain final forecasts. Conventional multi-layer perceptron (MLP) is also constructed for comparison purposes. The results show that both MTUNN and PENN outperform conventional MLP in the study. On average, MTUNN can obtain 8.1% improvement of MSE and 4.4% improvement of MAPE in comparison with MLP. PENN can achieve 10.5% improvement of MSE and 3.3% improvement of MAPE in comparison with MLP.

A multiple-path routing strategy for vehicle route guidance systems

Abstract This paper analyzes the transportation effectiveness of a multiple-path routing strategy using traffic simulation from the perspective of planning and designing a vehicle route guidance system. The test results indicate that the multiple-path routing strategy performs better than the commonly used shortest-path routing strategy.

Assessing Casualty Risk of Railroad-Grade Crossing Crashes Using Zero-Inflated Poisson Models

A railroad grade crossing (RGC) is a spatial location where rail and highway users share the right-of-way. A significant number of traffic crashes and severe consequences at RGCs have signaled the need for appropriate models to investigate the key factors associated with the casualty risk level at an RGC in terms of the number of fatalities or injuries caused by one or more crashes in a specific time period. This study used a zero-inflated Poisson regression model to describe the relationship between the extra-zero count fatality or injury data and explanatory variables collected at 592 RGCs in Taiwan. The annual averaged daily traffic and the presence of Guidance Sign 31 were significantly associated with the probability of no fatality or injury encountered at an RGC; if an RGC was at risk of a fatality or injury, the number of daily trains, crossing angle, and Guidance Sign 31 significantly influenced the expected total number of fatalities or injuries caused by traffic crashes. The empirical results indicated that traffic exposure and traffic signage have significant effects on the risk levels of casualties at an RGC.

Model crash frequency at highway–railroad grade crossings using negative binomial regression

Despite the fact that traffic collisions at highway–railroad grade crossings (HRGXs) are rare events, the impact of HRGX crashes is nevertheless more severe than highway crashes. Empirical studies show that traffic collisions at HRGXs are mainly attributed to railway-related and/or highway-related characteristics, particularly drivers’ abnormal behavior, driving around, or through an HRGX. These factors have different effects on crash likelihood (i.e., the number of traffic collisions or crash frequency) at an HRGX. To explore the causal relationship between crash frequency and the factors related to railroad and highway systems, we used a negative binomial regression model to identify the factors that are statistically significantly associated with traffic collisions at HRGXs, and conducted relevant sensitivity analyses to investigate the marginal effect of daily highway traffic on changes in crash frequency. The empirical study shows that the number of daily trains, the number of tracks, highway separation, annual averaged daily traffic (AADT), and crossing length had statistically significant effects on the mean number of traffic collisions (all p-values ≤ 0.0487). Further, the marginal effect of the AADT on the change of crash frequency indicates that crash likelihood monotonically increases with the increase of AADT.

The Analysis of Train Reliability for the Taiwan High Speed Rail

This paper on the analysis of train reliability for the Taiwan High Speed Rail (THSR) system is from the proceedings of the 12th International Conference on Computer System Design and Operation in Railways and Other Transit Systems, held in Beijing, China, in 2010. The authors report on a study that analyzed the service reliability of the Taiwan High Speed Rail in terms of punctuality and average delay per train. They discuss the concept of risk management as it is used to analyze the frequency and the severity of train delays caused by different kinds of accidents. The study found that signal and interlocking failures are the main reasons leading to train delays; earthquakes and typhoons are also major threats to the system, even though the system tends toward stability. The authors conclude by recommending shortening the maintenance cycle to help alleviate the problem of train delays caused by signal failures.

DEMAND-RESPONSIVE PRICING METHOD FOR THE PRODUCT LINE OF TAIWAN HIGH-SPEED RAIL

Presented is a demand-responsive pricing method for the product line of Taiwan high-speed rail. Considered are the objectives of both the operator and the passenger with their leader-follower and noncooperative relationships. The method is constructed by a bilevel program. A pricing model, which is the upper-level problem, considers three structures: basic model, market differential model, and product differential model. A demand model, which is the lower-level problem, considers service choice of passengers as a logit choice function where market competition is represented by a linear demand function and service capacity is used as a constraint for each service section. In a numerical study of three origin-destination pairs, presented are the results of the pricing models and their associated demand patterns. Convergence and sensitivity of model parameters are analyzed as well. The case study shows promising results for the flexibility and capability of the demand-responsive pricing method.

Using fault tree analysis to identify the failures of level crossing protection devices

Level crossings are potential hazardous locations in a railway system. To reduce the risk of level crossings, several active protection devices such as flashing lights and barriers may be installed. Previous studies have proved that such protection devices are effective in decreasing the probability of accidents. However, these protection devices may be partially out of order or receive error signals. Once protection devices cannot operate correctly, they may deliver wrong information that confuses the users. As a result, the users may trespass a level crossing in a dangerous situation. To realize the causes of different types of wrong operations, this study categorized them into error warning, conflict warning, and no warning. Then fault tree analysis is employed to identify the failure types and the causes behind the wrong operations of protection devices, including the failures of flashing lights, boom barriers, train direction indicators, obstacle detectors, and emergency buttons. The results indicate that error warning is mostly due to the error messages of the track circuit. Conflict warning is caused by specific failures of the protection devices. Finally, the wiring error of the track circuit is the main cause leading to no warning.