Re: how to solve this kind of optimization problem




In article <1193022469.259298.26040@xxxxxxxxxxxxxxxxxxxxxxxxxxx>,
Sam <hollowspook@xxxxxxxxx> writes:
Hi, there.

I wish I can find some help in this forum.

First I have an optimization problem as follows.

min f(x[1], x[2], x[3], x[4],..., x[99], x[100])
s.t. h(x[1], x[2], x[3], x[4],..., x[99], x[100]) = 0;
g-<g(x[1], x[2], x[3], x[4],..., x[99], x[100]) <g+;
x-<x[i]<x+, i =1, 2, 3,..., 100;

This can be solved by interior-point method.

However suppose that there is a new constraint on x now.
sum(k(x[i],x[i+1]))<=50; i = 1, 2, 3,...99;
where k(x[i],x[i+1])=0 if x[i]~=x[i+1] and k(x[i],x[i+1])=0 if
x[i]==x[i+1]

something is a misprint here
maybe you meant "=1" if x[i] not= x[i+1]

? no x[i] should currently change by more than 50 or the total sum?
anyway: a discontinuous constraint like your cannot be used within
nonlinear optimizeers which not only assume continuity but even
C2 - smoothness (otherwise they might terribly fail)
if you want to avoid a too large change in the current x[i]
- whatever its value might be in the given bounds - then you
could do this only by going into the code:
an trust region optimizer would allow this by setting the maximum
trust region radius -
or you could for example try
sqrt ( sum( (x[i]-x[i+1])^2 ) ) <= 50
this is also unsmooth by only on the diagonal in R^n, so its not quite
probable that this will have a bad influence



The physical meaning of this new constraint is that we expect the
change on x[i] is less than a constaint value, which is 50 in this
example.

How to deal with this new constraint? Can this new optimization
problem still be solved by interior point method? Please do me a
favour.

Thanks in advance.

.



Relevant Pages

  • Re: how to solve this kind of optimization problem
    ... Also the exact meaning of this is that we expect the number of changes ... this would mean a terrible combinatorical optimization problem ... >First I have an optimization problem as follows. ... >However suppose that there is a new constraint on x now. ...
    (sci.math.num-analysis)
  • Re: Matlab optimization question
    ... I need some help regarding optimization and data-fitting. ... Can I formulate this as an optimization problem and use something like ... how can I specify the constraint which involves one of the ... unknown variables? ...
    (comp.soft-sys.matlab)
  • Re: upper and lower bounds as constraints?
    ... the optimization is also over the weights mentioned in the beginning ... The objective function is a sum of quantum probabilities given as ... THere is 1 equality constraint of similar ... any advice would be appreciated. ...
    (comp.soft-sys.matlab)
  • Re: Scaling in fmincon()
    ... > I have been using the Matlab constrained optimization fucntion(). ... does anyone know if Matlab fmincon() does any scaling at all? ... I'll form two objectives, ... Iter F-count fconstraint Step-size derivative ...
    (comp.soft-sys.matlab)
  • Re: Numerical Recipes (Fortran) Usenet ??
    ... optimization, particularly when the desired solution happens to ... occur right on a constraint boundary, or at an intersection of ... It isn't a Fortran matter. ...
    (comp.lang.fortran)