Refined Path Planning for Emergency Rescue Vehicles on Congested Urban Arterial Roads via Reinforcement Learning Approach

Algorithm 1

Rescue path planning for emergency vehicles based on PERDQN.

(i)	Initialization: minibatch , step-size , replay period and size , exponents and , budget .
(ii)	Initialize experience replay memory
(iii)	Assign the starting position of the emergency vehicle to the initial state
(iv)	Observe and choose action
(v)	for  = 1 todo
(vi)	Observe
(vii)	Store driving experience in with priority
(viii)	ifdo
(ix)	fortodo
(x)	Sample driving experience
(xi)	Compute importance-sampling weight for experience
(xii)	Compute TD error based on equation (9)
(xiii)	Update experience priority
(xiv)	Accumulate weight-change
(xv)	end for
(xvi)	Update weights in Q-network according to equation (6) and then reset
(xvii)	Every steps copy weights into target network
(xviii)	end if
(xix)	With probability , choose action randomly
(xx)	Otherwise, choose action
	end for