Multitask Learning with Graph Neural Network for Travel Time Estimation

Algorithm 1

The dependency between road segments along the path.

	Input: sp, p, , ts, t
	Output:
	Initialize matrix: randomly
	Initialize vector: randomly
	Initialize scalar: randomly
	// sp: the speed tensor.
	// : the speed after GAT operation.
	// P: the given path.
	// : the neighbor road segments of the object road segment.
	// ts: the start time of the given path.
	// t: the return travel time of the road segment.
	// j: the ID number of the object road segment.
(1)	//The road segment ID involving the given path
(2)	while i < |p|
(3)	[i] = sp[i]
(4)	//The time step K;
(5)	for s = ts; s < K; s++ do
(6)	//The neighbor road ID of the object road segment.
(7)	for j = i; j < |N (i)| + 1; j++ do
(8)	// The correlation of the neighbor road segments.
(9)	(see (4))
(10)	[j][s] =
(11)	end for
(12)	[j][s + t] =
(13)	end for
(14)
(15)	end while
(16)	Return