- View SourceHi all,

As a part of academic project, I am solving one Cable

Laying problem using GP.

Problem is:

To optimise cost, minimum path of laying cable in area

X to be found.

Inputs are Road, Obstacle (bldg, pond, temple etc),

Delivery points, Source Points.

So, using two dimensional array I represented all

these points like

Path Path Path

Path Road Path

Source Road Path

Path Road Obstcale

Path Road Path(D)

Path(T)Path Path

T- Turning point (Assumption)

D- Delivery point

Now, to connect with Source to Delivery Point I can

create Turning Points to minimise distance in overall graph.

In my initial population I will have to create minimum

spanning trees of all possible graphs with Turning points.

My problem is:

How can I create graph by connecting edges where

Obstacle is present. Ofcourse, path is available

but lots of turning points may require. So, to walk

and find the path is very messy exercise.

And also how can I evaluate fitness? Because if i take

any formula like Eucilidean etc. then also it will be incorret due to

presence of obstacle between the points.

Pls reply soon.

-bhavin - View SourceI find your problem specification somewhat confusing. I would be

helped by a sample arrray of about 5 x 5 together with a text

describing what the array represents.

It might also be useful to know what resolution is desired. Some

methods that might be useful for a 100 x 100 array might not work well

due to space or time considerations on a 10,000 x 10,000 array.

Hope I can help.

george

--- In aima-talk@yahoogroups.com, "Bhavin" <bhavin.sanghani@g...> wrote:

> Hi all,

>

> As a part of academic project, I am solving one Cable

> Laying problem using GP.

>

> Problem is:

> To optimise cost, minimum path of laying cable in area

> X to be found.

>

> Inputs are Road, Obstacle (bldg, pond, temple etc),

> Delivery points, Source Points.

> So, using two dimensional array I represented all

> these points like

>

> Path Path Path

> Path Road Path

> Source Road Path

> Path Road Obstcale

> Path Road Path(D)

> Path(T)Path Path

>

> T- Turning point (Assumption)

> D- Delivery point

>

> Now, to connect with Source to Delivery Point I can

> create Turning Points to minimise distance in overall graph.

>

> In my initial population I will have to create minimum

> spanning trees of all possible graphs with Turning points.

>

> My problem is:

> How can I create graph by connecting edges where

> Obstacle is present. Ofcourse, path is available

> but lots of turning points may require. So, to walk

> and find the path is very messy exercise.

>

> And also how can I evaluate fitness? Because if i take

> any formula like Eucilidean etc. then also it will be incorret due to

> presence of obstacle between the points.

>

> Pls reply soon.

>

> -bhavin