Development of Travel Time Prediction Models Using Statistical and GEP Techniques for Public Transportation Bus Routes in Amman-Jordan

The efficiency and dependability of bus systems have an impact on commuters' daily experiences in today's urban transportation landscape. A key issue for transportation agencies is predicting bus travel times accurately as this directly affects the efficiency of the transit system. This research explores the world of predicting bus travel times with a focus on tackling the challenges that come with managing this aspect of public transportation.

The public transportation system in Amman currently grapples with a series of challenges, including a limited range of vehicle types such as minibuses, buses, and taxis. The private minibuses, predominantly organized by the Greater Municipality of Amman (GAM), face operational issues, notably the absence of guidelines and timetables. Operators prioritize profit over service quality, leading to concerns about demand, reliability, and efficiency. The absence of a timetable, poor management, random distribution of tracks, and a lack of future planning contribute to the system's inefficiency.

In the context of public transportation systems, predicting bus travel time poses a unique challenge due to several intricate factors that distinguish buses from passenger vehicles. Unlike cars, buses follow specific routes with designated stops, introducing complexities such as dwell time caused by boarding and alighting passengers. Acceleration and deceleration patterns for buses differ significantly from those of individual vehicles. The unpredictability of traffic volume, delays at traffic signs, and the random nature of stops, where buses halt based on passenger decisions rather than at predefined stops, further compound the difficulty of accurately forecasting travel time. Traditional approaches that focus on passenger vehicle travel time may not be directly applicable to buses, necessitating a specialized model tailored to the distinct behavioral and mechanical features of buses within the public transportation network. The research aims to contribute to the improvement of the public transit system in Amman by developing an accurate and reliable model for predicting bus travel time. Genetic Expression Programming (GEP) and SPSS are proposed as suitable tools for this task based on available historical data. And Statistical model which involves the use of mathematical models, methods, and tools to analyze and interpret data using SPSS program. The primary objectives include identifying the main factors influencing bus travel time, creating predictive models applicable to various bus routes, and establishing a model that considers the relationship between bus travel time and passenger vehicle travel time on the same routes.

In summary, the research endeavors to provide practical solutions to the challenges faced by the public transit system in Amman, with a focus on predictive modeling using GEP. The comprehensive approach considers various factors influencing travel time and emphasizes collaboration with stakeholders for effective implementation and improvement of the public transportation system.

Because of the congestion problem and to encourage passenger use of public transportations instead of the private vehicle, this study discussed the factors affecting the bus travel time that are operating with mixed traffic for a minibus in Amman and developed two models to predict bus operating speed by using SPSS and Gene Expression Programming (GEP). The model for predicting bus operating speed in mixed traffic lanes in Amman was developed using multiple regression analyses and genetic algorithm techniques with a 95% confidence level. The main factors affecting the bus travel time were the number of bus stops, the number of intersections, traffic volumes, and route length. These factors achieve a highly significant relationship with bus travel time.

The results of model 1 showed: the number of bus stops, signalized intersections, route length, and the average traffic volume are the most effective factors that affect Bus operating speed. Also, the predicted model has a high coefficient determination (R-square) with 0.888 and 0.93 for SPSS and GeneXpro5.0, respectively.

This result showed that the genetics algorithm gave a stronger relationship and more efficient model than multiple regression using SPSS. The results of Model 2 showed: the number of bus stops and the speed of the private vehicle have a strong relationship with the bus operating speed with the coefficient of determination (R-square) with 0.96 and 0.97 for SPSS and GeneXpro5.0, respectively. This result shows that there is no significant difference between the two techniques. Model 2 can help the planning engineers to predict bus operating speed for the new routes by knowing the private vehicle speed and the number of proposed bus stops along the new routes.

The following recommendations are depending on the observation of the transit system should be taken into consideration:

1-The authorities must enforce bus drivers to stop only in their certified locations for boarding and lighting passengers.

2-Greater Amman Municipality (GAM) should improve the public transportation system by studying the possibility of using an exclusive bus lane to attract more passengers and reduce the delay This achieved by collaboration between transport authorities and urban planners is essential. Identifying the main bus routes, conducting feasibility studies, and obtaining public support are important steps. In addition, flexible lane allocation strategies, such as dynamic bus lanes during peak hours, can also be explored to mitigate potential problems.

3-Using the bus with multi-door to reduce the dwelling time.

4-Improve the reliability and serviceability of the transit system.

5-Improve quality control over the bus operators.

6-To improve the transit system in Amman, future research should be taken into consideration and implemented, such as fuel consumption, cost of delay time, maintenance of exclusive lanes, and the effect of weather conditions on bus operating speed.

Studies on the impact of bus travel are crucial for understanding transportation efficiency, accessibility, and urban planning. They can improve efficiency by identifying bottlenecks and inefficiencies in bus routes, improving accessibility by identifying areas with limited access and addressing environmental impact by optimizing routes and infrastructure. Technological advancements like real-time tracking systems and predictive analytics can further refine bus travel times. Future studies should focus on assessing the effects of interventions, exploring the link between travel time variability and passenger contentment, evaluating alternative transportation modes, integrating social equity and environmental sustainability into transportation planning, and exploring the potential of emerging technologies like autonomous vehicles and electrification.

This research can be used for study the effect of bus stops specification and performance on the Bus operating speed and the Impacts of Bus Stops near Signalized Intersections on bus operating speed.

Although our study focuses on a specific geographic area and time period, we recognize the potential impact of changes in local transportation patterns. This realization prompted deep reflection on the generalizability of our findings. We encourage future research to explore the applicability of our results to different urban environments with different traffic dynamics, as this will provide a more comprehensive understanding of the factors influencing bus travel time predictions.

At its core, our methodology seeks to balance specificity and generality, laying the foundations for developments that adapt to different urban landscapes and transport systems.