TY - JOUR
T1 - Prescriptive Analytics for Intelligent Transportation Systems with Uncertain Demand
AU - Wang, Huiwen
AU - Yi, Wen
AU - Tian, Xuecheng
AU - Zhen, Lu
N1 - Funding Information:
This research was supported by the National Natural Science Foundation of China (Grant Nos. 72201229 and 72361137006). All authors contributed equally to the work. All authors contributed equally to the work and are co-first author.
Publisher Copyright:
© 2023 American Society of Civil Engineers.
PY - 2023/12/1
Y1 - 2023/12/1
N2 - Data-driven traffic modeling is revolutionizing transportation systems and provides numerous opportunities for achieving high-quality transportation services. A major challenge in optimizing transportation systems is uncertain transportation demand. With the availability of historical data on transportation demand, the uncertain transportation demand can be better modeled, and thereby practitioners can formulate well-informed transportation scheduling decisions. In this paper, we propose three effective and economical transport scheduling strategies using mathematical programming, leveraging big data to extract useful contextual information. Additionally, a perfect-foresight optimization model is proposed to evaluate our proposed data-driven strategies. Results show a negligible optimality gap (i.e., 0.47%) between the optimal solution derived by the perfect-foresight model and the scheduling plans derived by our data-driven strategies. Overall, this paper contributes to the field of transportation engineering by innovatively applying data science, mathematical modeling, and optimization techniques.
AB - Data-driven traffic modeling is revolutionizing transportation systems and provides numerous opportunities for achieving high-quality transportation services. A major challenge in optimizing transportation systems is uncertain transportation demand. With the availability of historical data on transportation demand, the uncertain transportation demand can be better modeled, and thereby practitioners can formulate well-informed transportation scheduling decisions. In this paper, we propose three effective and economical transport scheduling strategies using mathematical programming, leveraging big data to extract useful contextual information. Additionally, a perfect-foresight optimization model is proposed to evaluate our proposed data-driven strategies. Results show a negligible optimality gap (i.e., 0.47%) between the optimal solution derived by the perfect-foresight model and the scheduling plans derived by our data-driven strategies. Overall, this paper contributes to the field of transportation engineering by innovatively applying data science, mathematical modeling, and optimization techniques.
KW - Data-driven transportation modeling
KW - Large-scale optimization
KW - Prescriptive analytics
KW - Uncertain demand
UR - http://www.scopus.com/inward/record.url?scp=85173869988&partnerID=8YFLogxK
U2 - 10.1061/JTEPBS.TEENG-8068
DO - 10.1061/JTEPBS.TEENG-8068
M3 - Journal article
AN - SCOPUS:85173869988
SN - 2473-2907
VL - 149
JO - Journal of Transportation Engineering Part A: Systems
JF - Journal of Transportation Engineering Part A: Systems
IS - 12
M1 - 04023118
ER -