11.12.2012 Views

Chapter 5 Robust Performance Tailoring with Tuning - SSL - MIT

Chapter 5 Robust Performance Tailoring with Tuning - SSL - MIT

Chapter 5 Robust Performance Tailoring with Tuning - SSL - MIT

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Data: initial iterate, x0, initial temperature, T0, change bounds on x, ∆x,<br />

algorithm parameters, ka, kr, kc, ndof<br />

Result: optimal design variables and cost: x ∗ , J ∗<br />

begin<br />

Evaluate initial cost: J0 = f (x0)<br />

Initialize: x = x0, J = J0, k =1,T = T0, A = R =0<br />

while not frozen and number of function evaluations < max do<br />

Randomly choose design variables to change based on ndof<br />

Generate random change <strong>with</strong> bounds ∆x [Equations 2.42 and 2.43]<br />

if xk is feasible then<br />

Evaluate new cost, Jk = f (xk)<br />

Calculate change in cost, ∆J = Jk − Jk−1<br />

if ∆J kr or A>kA then<br />

Increment temperature [Equation 2.45]<br />

Reset counters, A = R =0<br />

end<br />

end<br />

end<br />

Return lowest-cost design: J ∗ =minJ, x ∗ =argminJ (x)<br />

end<br />

Figure 2-4: Simulated annealing algorithm.<br />

58

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!