sci.op-research

Hi Christophe,

Thank you for your reaction on my questions.

With regards to optimality or sub-optimality I dont believe that this is
necessary. If there is a rather good solution to this problem which is not
optimal than it would also be satisfactory.

I didn't look at nurse rostering untill now. I will give a good look at
articles concerning this type of problem.

The problem size would be fairly small. I still don't know the exact size,
but I will probably know it in the next week when I have a conversation with
the planners. Then I will be able to know for sure the complexity of this
problem. I will make a post when I know more about the problem size.

Thanx again for your input.


christophe.vanhuele@gmail.com wrote:
>> Dear everybody,
>>
>[quoted text clipped - 59 lines]
>> Business Mathematics and Informatics
>> Vrije Universiteit Amsterdam
>
>Hi,
>I'm only a student Operational Research at Ghent University,
>but I'll give it a try:
>Before you implement your heuristics/algorithm, make sure that the
>complexity of the problem isn't too high.
>Try to 'evaluate' the number of possibilities.
>If these are exponential, I would opt for a meta-heuristic instead of
>exact procedures.
>That would save you time in algorithm design.
>
>And second, there is definitely literature for this problem:
>you should look at nurse rostering in academic journals (EJOR).
>
>And most importantly, do you need an optimal or a near-optimal
>solution?

--
Message posted via MathKB.com
http://www.mathkb.com/Uwe/Forums.aspx/op-research/200804/1