You may want to look at our CSES portfolio. It can be applied to your question with respect to risk management. One of our papers reference Barry Boehm's work in this area. http://academic.research.microsoft.com/Paper/4813650.aspx

Our work:

DOI: 10.1109/HICSS.2009.308

DOI: 10.1109/CICYBS.2009.4925092

All the best.

Bob

Thank you Sir,

But is this problem is directly mentioned in any paper or has been proved?

Professor Chris Verhoef has done a lot of research in this area. You can view his papers on this and related subjects on http://www.cs.vu.nl/~x/qipm.html.

Thank you Sir, Mr. Frank Vogelezang,

Sir, i have searched many papers bt didnt get any exact reference regarding software cost estimation before requirement collection is a problem.

So, i need atleast one exact reference in this area.

Hello! Does this question have been solved? It looks like interesting such hypothesis, a priori I guess that software cost estimation is impossible without requirements, how can you define a cost if you don´t know what are you going to do?