Development of trip generation and attraction mdels For Hotel Facilities in Kampala

Abstract

Trip generation/attraction models have been widely applied in transport planning and traffic control. The models relate facility attributes and adjacent road network traffic characteristics to predict future trip attractions or generation from a facility. The aim of the study was to (i) establish factors that significantly affect/influence traffic trip attraction and generation for Hotel developments, (ii) derivate peak trip attraction and generation traffic from traffic counts conducted & (iii) to conduct sensitivity analysis so as to understand the gap between the conceptual and actual implementation of the models. According to literature, most models are based on empirical studies conducted in the United States, Uk etc however, differences in land-use type, facility design, adjacent road network design, and mixed-use characteristics of the facilities create limitations on transferability of the models which affects reliability and therefore accuracy of prediction. The trip attraction and generation models were estimated for 3- and 4-star Hotels that account for approximately 50% of the total classified Hotels in Kampala City. Model development was conducted using negative binomial regression with the dependent variable being number of trips in and out of the hotel facility in the peak hour which is a discrete random variable and number of hotel rooms, conference facility size, number of parking slots and peak hour traffic in both directions on the adjacent road being the explanatory variables. Model selection was based on goodness-of-fit tests namely; lower Akaike Information criteria (AIC), likelihood ratio statistic, and the likelihood ratio tests in R software. The estimated models for trip attraction and generation were significant at 95% confidence levels with R2- values of 67.7% and 68% respectively. The developed models where based on the relationship between selected Hotel attributes and traffic generated/attracted to Hotels, however other intervening variables such as proximity of the hotel to the Central Business District (CBD), weather pattern (dry/wet days), traffic behavior during special congregational functions at hotels or in the vicinity, traffic behavior while VIPs are using the roads etc should be investigated to estimate trip attraction and trip generation models. The resulting models should then be compared with the models in the current study.