Trajectory Optimization of CAVs in Freeway Work Zone considering Car-Following Behaviors Using Online Multiagent Reinforcement Learning

Table 1

Parameters and values.

Parameter	Value
Maximum deceleration	−6 (m/s^2)
Accepted deceleration	−2.5 (m/s^2)
−1 m/s^2 per distance	200 (m)
Average standstill distance	1.12 (m)
Safety distance reduction factor	0.60
Additive part of safety distance	1.85
Multiplic. part of safety distance	3.02
Learning rate	0.1
Discount factor	0.9
Exploration rate	0.2