Prev: struct_ array
Next: MatLab GUI, image
From: zac gross on 29 Jul 2010 01:05
Hi,

I want to be able to optimize a time series set of data, such that it is bound by both linear inequalities and such that the rate at which it can change is constrained.

ie such that

X + Y + Z > Demand for all t
and that
X(t) - X(t-1) < 5 and similarly for Y and Z


While trying to minimize a total cost function along the lines of

total cost = x*sum(X) + y*sum(Y) + z*sum(Z)

Where x,y,z are the respective average costs of X,Y,Z

Can someone suggest what program/function I should use?

Thanks
From: zac gross on 30 Jul 2010 01:14
"zac gross" <i.gross(a)ugrad.unimelb.edu.au> wrote in message <i2r260$7uf$1(a)fred.mathworks.com>...
> Hi,
>
> I want to be able to optimize a time series set of data, such that it is bound by both linear inequalities and such that the rate at which it can change is constrained.
>
> ie such that
>
> X + Y + Z > Demand for all t
> and that
> X(t) - X(t-1) < 5 and similarly for Y and Z
>
>
> While trying to minimize a total cost function along the lines of
>
> total cost = x*sum(X) + y*sum(Y) + z*sum(Z)
>
> Where x,y,z are the respective average costs of X,Y,Z
>
> Can someone suggest what program/function I should use?
>
> Thanks



Would I have to separate every data point into a separate variable?

How can I define the Objective Function with vectors instead of scalar variables?
From: Alan Weiss on 30 Jul 2010 08:33
On 7/30/2010 1:14 AM, zac gross wrote:
> "zac gross" <i.gross(a)ugrad.unimelb.edu.au> wrote in message
> <i2r260$7uf$1(a)fred.mathworks.com>...
>> Hi,
>>
>> I want to be able to optimize a time series set of data, such that it
>> is bound by both linear inequalities and such that the rate at which
>> it can change is constrained.
>>
>> ie such that
>> X + Y + Z > Demand for all t and that X(t) - X(t-1) < 5 and similarly
>> for Y and Z
>>
>>
>> While trying to minimize a total cost function along the lines of
>> total cost = x*sum(X) + y*sum(Y) + z*sum(Z)
>>
>> Where x,y,z are the respective average costs of X,Y,Z
>>
>> Can someone suggest what program/function I should use?
>>
>> Thanks
>
>
>
> Would I have to separate every data point into a separate variable?
>
> How can I define the Objective Function with vectors instead of scalar
> variables?

Take a look at Optimization Toolbox documentation:
http://www.mathworks.com/access/helpdesk/help/toolbox/optim/ug/brhkghv-18.html#brhkghv-21
shows what kinds of problems the toolbox addresses. Most solvers handle
multidimensional problems.

The toolbox documentation is arranged along the usual workflow:
1. Decide on a solver
2. Write objective and constraint functions
3. Set options
4. Run the solver
5. Examine results
6. Take appropriate steps

There is a section describing the Optimization Tool GUI, and one that
gives the underlying algorithms and associated examples.

Good luck,

Alan Weiss
MATLAB mathematical toolbox documentation
 | 
Pages: 1
Prev: struct_ array
Next: MatLab GUI, image