This preview shows page 1. Sign up to view the full content.
Unformatted text preview: permuta1on: • Applying a Swap Operator to a permuta1on: – (a, d, g, b, e) + SO(1,2) = (d, a, g, b, e) • Apply a sequence of SO (SS) to a permuta1on: – SS=(SO(1,2,), SO(5,4), SO(5,1)) – (a, d, g, b, e) + SS
> (d,a,g,b,e)
> (d,a,g,e,b)
> (b,a,g,e,d) Subtract Two Permuta1ons • A: (a, c, d, e, b), B: (c, a, b, e, d) • There is a SS that transforms A to B. – a is in posi1on 1 in A and 2 in B: SO1(1,2) – A+SO1=A’= (c,a,d,e,b) – b is in posi1on 5 in A’ and 3 in B: SO2(5,3) – A’+SO2=(c,a,b,e,d)=B • SS=(SO1(1,2), SO2(5,3)) • A
B = SS 4 2/25/14 TSP
PSO algorithm Par1cle Velocity Update
Con1nued Each velocity is a swap sequence (SS). v(t) = v(t
1) α*(pbest
x(t
1)) β*(gbest
x(t
1)) α, β are probability between 0 and 1. α * (pbest
x(t
1)) means when a generated random number is > α , all swap operators in swap sequence (pbest – x(t
1)) are included in the updated velocity. The same applies to β and (gbest
x(t
1)). • Other work has improved this simple scheme. •
•
•
• Exercise a
b
c
d
e
f x
16.47
16.47
20....
View
Full
Document
This document was uploaded on 02/28/2014 for the course CS 6776 at Memorial University.
 Winter '14
 DrYu

Click to edit the document details