An Optimal Composition Strategy for Knowledge Service Component Based on Flexible Tracking Particle Swarm Algorithm

Algorithm 1

	Input: the requirement attribute spaces and resource service spaces , variability of requirement attributes .
	Output: the component sequence of different DKEij.
	Step 1: set , ;
	Step 2: , ;
	Step 3: do {
	Searching for the variable threshold to meet min
	If ()
	{
	Incorporate and its child attributes into sd;
	For ()
	{
	Set  = 1, that is, fi has been mapped as an independent component or part of a component;
	Go to step 4;
	}
	}
	Else if ()
	{
	Set , that is, fi cannot be directly mapped as a component;
	Go to step 3;
	}
	} until (nonexistent such )
	Go to step 5;
	Step 4: if ()
	{
	All the characteristic of sd could be encapsulted into the different granularity components with the same functionality but different performance.
	Join into ;
	Set , go to step 3;
	}
	Else
	Go to step 3;
	Step 5: for ()
	{
	Map f into , and add into ;
	Set  = 1;
	}
	Step 6: output ;
	If ( = )
	{
	Develop again;
	}