Online Program

Return to main conference page

All Times ET

Program is Subject to Change

Tuesday, June 15
Tue, Jun 15, 9:30 AM - 11:00 AM
TBD
More Data, Geography, and Reduce Burden? Approaches to Improve Measurement of Goods Movements in the United States

Exploratory Research on Optimization of CFS Sampling Design (307992)

*Mehdi Hashemipour, US DOT 

Keywords: Sampling Design, Optimal Stratification and Allocation, Commodity Flow Survey, Genetic Algorithm

The term ‘sampling’ has been used to refer to situations in which part of a population is taken to be representative of the entire population. In other words, sampling is the process of observing selected members to approximate characteristics of the whole population from which they are drawn. Sampling design is defined as the methodology by which the sample units are chosen. With the advancement in computational power, the stratification and allocation is performed in one phase using a technique called joint stratification-allocation method. In this exploratory research, researchers used the R SamplingStrata package that applies the optimal stratification and allocation method based on Genetic Algorithm and Simulated Annealing in a CFS-like scenario. With this method, we aim at minimizing the total sample cost while satisfying the target Coefficient of Variation constraints. Experimenting these methods by using variables such as geographies, measure of size, and industry may result in different designs for the CFS sample.