Mei-Lam Tam

Reliability analysis of traffic noise estimates in Hong Kong

Abstract This paper examines the reliability of traffic noise measurement techniques and the noise estimates in Hong Kong. A simulation model making use of the Monte Carlo technique is devised to incorporate the uncertainty in traffic noise estimates in Hong Kong. Traffic flow, traffic speed and traffic composition (in terms of percentage of heavy vehicles) are identified as the key factors influencing the generation of traffic noise. The new equations for estimating the noise descriptors L10, Leq, L50 and L90 were calibrated on the basis of the Hong Kong survey results. The effects of the key factors on traffic noise estimates can be assessed. The probability distribution for each of the key factors is derived with the use of the survey data. The variation on the key factors is due to the sampling error in the survey. The reliability of the traffic noise estimates is obtained from the combined probability distributions of the key factors. The reliability analysis gives the traffic noise estimates with a particular probability or vice versa.

Reliability of territory-wide car ownership estimates in Hong Kong

This paper presents an aggregate car ownership model for Hong Kong and examines the reliability of territory-wide car ownership estimates from the model. Car ownership is defined as the total number of licensed private cars and motorcycles (in terms of passenger car units). The probability distributions for the relative errors of actual and projected values of the key factors were utilized to explain the variation of the future values of these factors. A simulation model making use of the Monte-Carlo simulation technique was devised to incorporate uncertainty in the estimation of territory-wide car ownership in Hong Kong, in which the reliability of territory-wide car ownership estimates was obtained from the combined probability distribution of the key factors. The reliability analysis produces territory-wide car ownership estimates at a particular probability, or probabilities of reaching specific car ownership levels.

ANALYSIS OF DEMAND FOR ROAD-BASED TRANSPORT FACILITIES: A BI-LEVEL PROGRAMMING APPROACH

A bi-level programming model is presented for determining the number of parking spaces and road-capacity expansions that are required to satisfy road traffic and parking demand. The lower-level problem is an equilibrium trip-distribution/assignment problem, and the upper-level problem is to maximize the net user benefit (or consumer surplus) by considering travelers’ route and destination choice behavior and by satisfying the network capacity and attraction-end parking supply constraints. The use of consumer surplus in the objective function reflects an economic tradition that public investment in transport facilities (parking spaces and road capacities) should be operated in a way that maximizes the facilities’ social benefits. A sensitivity-analysis-based heuristic algorithm is developed to solve the proposed bi-level problem and is illustrated with numerical examples.

Why standard modelling and evaluation procedures are inadequate for assessing traffic congestion measures

As traffic congestion in Hong Kong worsens with the growing use of motor vehicles, to determine what has to be done to achieve and maintain an acceptable level of mobility for persons and goods becomes the subject of intense debate by economists, planners and politicians. Although many traffic congestion measures have been suggested and investigated, they are mainly confined to expansion of transport infrastructure and measures to make effective use of roads. New roads and public transport improvements may encourage increases in traffic and changes in the patterns of trips, while traffic management and regulation measures may in turn lead to suppression in road traffic. In order to evaluate these measures, it is required to make use of a transport model in assessing their traffic impacts. However, there are always discrepancies between the traffic forecasts and the actual flows on the roads. This paper investigates why the standard modelling and evaluation procedures currently used by the Hong Kong Government are inadequate for assessing the traffic congestion measures. Empirical evidence is given together with discussion on modelling and evaluation issues raised by the existence of suppressed/induced traffic.