TY - JOUR

T1 - A heuristic-based population synthesis method for micro-simulation in transportation

AU - Zhuge, Chengxiang

AU - Li, Xia

AU - Ku, Chia An

AU - Gao, Jian

AU - Zhang, Hui

N1 - Publisher Copyright:
© 2017, Korean Society of Civil Engineers and Springer-Verlag Berlin Heidelberg.

PY - 2017/9/1

Y1 - 2017/9/1

N2 - Population synthesis is extensively required by a number of micro-simulation models in transportation. A heuristic-based population synthesis method called Pop-H was proposed to overcome the following two limitations that received less attention. The first limitation is that one target marginal distribution can be well met by various sets of household weights that can be used to generate different sets of population and thus it is a problem that which set of household weights is the real one. Secondly, the population synthesis is commonly viewed as an optimization problem, and minimizing the Mean Absolute Percentage Error of control variables is generally used as the objective function. The Standard Deviation of control variables is also crucial in some cases, which, however receives scant attention. In response to these two limitations, the heuristic-based population synthesis method works in the following way: the Pop-H algorithm starts with the initial set of household weights derived from a sample data and calculates the final set of household weights by iteratively adjusting the initial set in a defined way with the objective function taking into account both Mean Absolute Percentage Error and Standard Deviation of control variables. Finally, the medium-sized city of Baoding, China was used as the case study. The sensitivity test was firstly done to examine four key parameters of the Pop-H algorithm, and then the algorithm was applied to create the population for the whole city.

AB - Population synthesis is extensively required by a number of micro-simulation models in transportation. A heuristic-based population synthesis method called Pop-H was proposed to overcome the following two limitations that received less attention. The first limitation is that one target marginal distribution can be well met by various sets of household weights that can be used to generate different sets of population and thus it is a problem that which set of household weights is the real one. Secondly, the population synthesis is commonly viewed as an optimization problem, and minimizing the Mean Absolute Percentage Error of control variables is generally used as the objective function. The Standard Deviation of control variables is also crucial in some cases, which, however receives scant attention. In response to these two limitations, the heuristic-based population synthesis method works in the following way: the Pop-H algorithm starts with the initial set of household weights derived from a sample data and calculates the final set of household weights by iteratively adjusting the initial set in a defined way with the objective function taking into account both Mean Absolute Percentage Error and Standard Deviation of control variables. Finally, the medium-sized city of Baoding, China was used as the case study. The sensitivity test was firstly done to examine four key parameters of the Pop-H algorithm, and then the algorithm was applied to create the population for the whole city.

UR - http://www.scopus.com/inward/record.url?scp=85007489171&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85007489171&partnerID=8YFLogxK

U2 - 10.1007/s12205-016-0704-1

DO - 10.1007/s12205-016-0704-1

M3 - Article

AN - SCOPUS:85007489171

SN - 1226-7988

VL - 21

SP - 2373

EP - 2383

JO - KSCE Journal of Civil Engineering

JF - KSCE Journal of Civil Engineering

IS - 6

ER -