Experiment design for determining the minimum sample size for developing sample based trip length distribution

Abstract

In transportation modelling, the origin and destination matrix is essential. To develop an origin-destination matrix, trip length distribution pattern is needed. Thus, it is important to develop a representative trip length distribution based on data sample. This paper present an attempt to develop experiment design for investigating the behaviour of curve pattern acceptance and error value acceptance due to the sample size variation. The experiment design indicate that, in general, smaller sample size produce higher error value. The experiment also indicate that even if the curve pattern is consider accepted, the error value can be unacceptable. Since the experiment design development is merely based on a particular small sample size, the research result can only give an indication yet the calculation result value can not be taken as a proper dictum. The experiment design can be considered as appropriate to be used for any further assessment.