Re: Military logistic problem



Unfortunately you are not very descriptive and give few elements about
the problem you have to adress.I believe that you problem may lie in
the field of Operational Research and belongs also in CS,because is a
programming(e.x linear,quadratic,etc).In case your problem can be
formulated as a set of linear inequation(linear programming),you can
apply a variety of techniques to find solution,hopefully enough you
problem can be solved in polynomial time.In case you problem is a
variant of linear programming,which is called integer programming,then
it is NP complete(the reduction is quite easy from the 3sat),so you
need exponential amount to solve it,on the other hand in case you don;t
need the optimal solution,then there are a quite well established
approximation techniques such as simulatead annealing,A* search
algorithms,randomized versions etc.some of them are quite robust and
can give you an index for the approximation ratio.
Please be more specific with your description.In case you need
something,you can email me.
Cheers
Yannis

.



Relevant Pages

  • Re: Scaling data
    ... about 90-95% of the data lie between 980-1000. ... "as close to something linear as possible". ... other 980 numbers (a natural average of about 10 points per identical bin). ... 980 maps to 140 ...
    (comp.soft-sys.matlab)
  • Scaling data
    ... I have about 100K data points that I'd like to rescale. ... And they are almost ramp-like, ... about 90-95% of the data lie between 980-1000. ... I'd like to rescale it to make it look as close to something linear as ...
    (comp.soft-sys.matlab)