A Composite Service Provisioning Mechanism in Edge Computing

Algorithm 1

The algorithm for the first users to seek optimized intermediate solution.

(1)	Input: the QoS of component service after the deployment location changes
(2)	Output: the optimized intermediate solution obtained in limited time
(3)	;
(4)	;
(5)	; \\ 0th generation probability model of three sub-populations are initialized.
(6)	= 0; \\The variable is initialized to zero.
(7)	The number of the superior sub-population is . The number of general sub-population is . The number of poor sub-population is ;
(8)	Three sub-populations are obtained according to their models sampling, respectively. The fitness value of three sub-population individuals are calculated, and sorted together by descending order. The top individuals are selected to form the superior solutions sub-population . The middle individuals are selected to form the general solutions sub-population . The last individuals are selected to form the poor solutions sub-population ;
(9)	while limited time do
(10)	;
(11)	For , select the top individuals form the group , and then update by the rules of statistics, smoothing and adding forgetting factor;
(12)	For , select the last individuals form the group , and then update by the rules of statistics, smoothing and adding forgetting factor;
(13)	First select some individuals with poor fitness values from the superior sub-population , then select some individuals with superior fitness values from the poor sub-population , finally form the general sub-population with the individuals in the general sub-population . Update according to formula (5);
(14)	Three new sub-populations are obtained according to their updated models sampling, respectively. The fitness value of three sub-population individuals are calculated, and sorted together by descending order. The top individuals are selected to form the +1th superior solutions sub-population . The middle individuals are selected to form the +1th general solutions sub-population . The last individuals are selected to form the +1th poor solutions sub-population ;
(15)	end while
(16)	Save the optimized intermediate solution in memory space;