Sensitivity analysis

[STM]

Hello,

I would like to perform standard sensitivity analyses by e.g. setting caps on total or sectoral GHG emissions. I understand this is possible only under the stochastic mode; now, I must say I do not really know how this could be formulated under VEDA-FE...

Any help is welcome !

Thanks,

Stephane

[Antti-L]

Well, I think it is strange if doing sensitivity analysis under VEDA-FE seems possible only under the stochastic mode.  But I guess Amit would be the best person to answer that question. I am quite sure that Amit has performed numerous sensitivity analyses with VEDA.

Doing standard sensitivity analysis (post-optimality analysis) for a linear programming model is always possible to the extent the solver supports it. You only need to specify the appropriate options in the solver options file. For more information on doing standard sensitivity analysis under GAMS, see http://support.gams-software.com/doku.php?id=solver: sensitivity_analysis_with_gams_cplex

For non-standard sensitivity analysis, the TIMES stochastic mode does, indeed, provide some tools for making the analyses faster and easier. To the extent that VEDA-FE supports all the stochastic parameters, setting up the runs should be rather straightforward. You basically only need to specify the number of "states of the world" and the "uncertain" bounds for the GHG emissions, and invoke the run with either the Stochastic or Sensitivity analysis mode activated.

However, I think one can equally well use the powerful scenario facilities of VEDA-FE for any non-standard sensitivity analysis, and that approach may appear to be the easiest solution for many users.

[STM]

Thanks for the answer Antti,

I can indeed manage the non-standard sensitivity analysis through scenarios (that's actually what I do for now).
Still, I'm afraid the VEDA-FE formulation of SOWs and uncertain bounds is not so straightforward for me... Any tip?

Thanks

[wnijs]

Following link might be helpful to better understand stochastic and SOW concepts:

http://www.etsap.org/docs/TIMES-Stochastic.pdf