结合粒子群优化算法

each particle has three elements : location, velocity and fitness. Loc has p number of directions. P = {a1,a2, .......a_p}

Sum{square a_i } =1

fitness = Q(a)

Pid[i] = (loc, fitness)： the optimized position and fitness for each particle

Pgd : the most optimized situation for the group of particle.

Particles[i].velocity(t+1) = w particle[i].velocity(t) + m1r1(Pid[i].loc(t) - Particle[i].loc(t)) +m2r2(Pid[i].loc(t) - Particle[i].loc(t)) 1

Particles[i].loc(t+1) = w particle[i].loc(t) + m1r1(Pid[i].loc(t) - Particle[i].loc(t)) +m2r2(Pid[i].loc(t) - Particle[i].loc(t)) 2

Step one : M number of particles,  the max interation Max gen R =p error

=e

Set M and max gen and R

And e

Step two:

Choose a as initial solution

Step three:

Calculate each fitness and set initial Verticity as 0

Step four:

Find the best Pid[i] and Pgd

Step five

Gen =1

Step six

Sumabs[(zt(i) - zt-1(I))]>=e and gen

Otherwise 14

Step seven

Calculate velocity and location according to 1 and 2 equation.

Step eight

Get the 投影值

Step nine

Get Q(a)

Step ten

Update Pid[i] and the most optimized Pgd;

Step eleven

Gen = gen+1

Step 12

Return to six

Step 13

Best a star and z star

Step 14

根据一惟K均值聚类算法、对最佳投影值zstar进行聚类结果分析