TY - JOUR
PY - 2023//
TI - Biobjective robust network-wide traffic signal optimization against cyber-attacks
JO - Transportation research part C: emerging technologies
A1 - Bao, Ji
A1 - Zheng, Liang
A1 - Ban, Xuegang (Jeff)
SP - e104124
EP - e104124
VL - 151
IS -
N2 - In the cyber age, modern traffic signal control systems are inevitably exposed to a wide range of cyber-attacks, which would negatively impact traffic system performance. To defend against direct cyber-attacks that alter signal settings directly and maliciously by hacking into the wireless connected traffic control center or signal controllers, this study applies a precautionary and inexpensive measure of designing the robust signal control plan. This is realized by modelling the biobjective traffic signal control simulation-based optimization (SO) problem under cyber-attacks in a bi-level framework, with the aim of minimizing the negative impact of potential cyber-attacks on traffic efficiency and safety performance. Depending on the type of attacker with complete or incomplete knowledge on traffic signal control system, two typical bi-level models exist. Then, we propose an improved biobjective robust simulation-based optimization (IBORSO) algorithm to solve the two models, during which the sequential interaction between the attacker and the defender is seen as the Stackelberg game, and the resulting equilibrium solutions are the desired robust signal control plans to withstand cyber-attacks. An urban road network in Changsha, China, is modelled as the experimental site and numerical results show the effectiveness of IBORSO in solving the two bi-level models. Meanwhile, the solved robust signal control plans can defend against two typical attacks satisfactorily and achieve better or comparable biobjective performance compared with the field-implemented one and optimized ones by biobjective efficient global optimization (BOEGO). Language: en
LA - en
SN - 0968-090X
UR - http://dx.doi.org/10.1016/j.trc.2023.104124
ID - ref1
ER -