Improving POI Recommendation via Dynamic Tensor Completion

Algorithm 2

The procedure of PIO.

	Input: pigeon location , velocity , map compass operator R, max iterations , , landmark , existing data , quality function ;
	Output: optimal location ;
1	initialize , , ,
2	while reach do
3	Calculate the velocity of each pigeon
4	Update the location of each pigeon
5
6	while reach do
7	Calculate landmark
8	Update
9	Calculate the location
10
11
12	return ;