Hi together, find attached a first draft for the tariff selection method. It strikes a balance between simplicity and economic meaningfulness. Cheers Chris 
Christoph Michael Flath Information Process Engineering (IPE) Tel: +497219654828 Fax: +497219654829 ============================================================================== FZI Forschungszentrum Informatik an der Universität Karlsruhe HaidundNeuStr. 1014, D76131 Karlsruhe Tel.: +4972196540, Fax: +497219654959 Stiftung des bürgerlichen Rechts Stiftung Az: 140563.1 Regierungspräsidium Karlsruhe Vorstand: Dipl. Wi.Ing.
Michael Flor, Prof. Dr. rer. nat. Ralf Reussner, Prof. Dr. rer. nat. Dr. h.c. Wolffried Stucky, Prof. Dr. rer. nat. Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
============================================================================== Tariff selection.pdf (725K) Download Attachment 
Administrator

Hi Chris, For some reason I get an error message when trying to open it. Thanks, Wolf From: chris.flath [via Power TAC Developers] [mailto:[hidden email]] Hi together, find attached a first draft for the tariff selection method. It strikes a balance between simplicity and economic meaningfulness. Cheers Chris  Christoph Michael Flath Information Process Engineering (IPE) Tel: +497219654828 Fax: +497219654829 ============================================================================== FZI Forschungszentrum Informatik an der Universität Karlsruhe HaidundNeuStr. 1014, D76131 Karlsruhe Tel.: +4972196540, Fax: +497219654959 Stiftung des bürgerlichen Rechts Stiftung Az: 140563.1 Regierungspräsidium Karlsruhe Vorstand: Dipl. Wi.Ing. Michael Flor, Prof. Dr. rer. nat. Ralf Reussner, Prof. Dr. rer. nat. Dr. h.c. Wolffried Stucky, Prof. Dr. rer. nat. Rudi Studer Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus ==============================================================================
If you reply to this email, your message will be added to the discussion below: http://powertacdevelopers.975333.n3.nabble.com/Tariffselectionproblemtp2833549p2833549.html To start a new topic under Power TAC Developers, email [hidden email] 
Apparently Nabble does not properly link the attachment in the emails sent out  works allright in the web interface.
Here is a direct link: Download 
Just noticed that there was a mathematical error in the document  for the quantal response equilibrium to work utility values need to be positive.

Administrator

Chris 
This is good, but it's missing a couple of important elements. Choosing a tariff is not like choosing a new bicycle. Everyone already has a tariff, so we need to model the utility of doing nothing, and only switch if the utility of some tariff is higher than the utility of doing nothing. The utility of doing nothing is not the utility of the current tariff, it's higher than that, and it's probably quite variable among a real population. The biggest part of the utility of doing nothing is convenience. This is why we are using a mailbox full of junk mail to represent the tariff market in the spec. For a customer to switch tariffs, there needs to be enough improvement in utility to overcome some sort of intertia. It's probably the same math as the tunneling probability for an elementary particle. It's affected by energy gradient and the height of the barrier. Another factor is the disutility of variable or timeofuse prices. In order to benefit from such a tariff, either the customer's (net) usage pattern needs to produce a good average price, or the customer must change their habits and shift some loads. Some customers will presumably have or invest in automation for this, and some will have to pay attention and regulate their energy use. Does this make sense? Thanks for working on this. You will need to integrate your text and math into the game spec quite soon  we need to put out a draft by the end of April. John 
Administrator

I almost forgot  you also need to include the signup bonus and earlywithdrawal penalty, as well as the possibility of an earlywithdrawal penalty for the existing tariff.
Also, this treatment applies to individual customers, but our models are population models. So your model needs to produce not a switch/noswitch choice, but rather the number of customers within the population that will switch. John 
In reply to this post by grampajohn
This is already adressed in the current description. The cost subutility is determined by sampling random consumption stretches and the customer than adjusting its consumption as much as he is willing to to the tariff properties  the hereby realized average cost determines the cost subutility. Clearly, the cost subutility will suffer if the tariff is not suitable for my consumption profile. The cost subutility is accompanied by the risk subutility which is determined by the variance of the realized cost values under this tariff and furthermore the reliability of a variable tariff, i.e. how variable has it been so far. 
In reply to this post by grampajohn
Yes  this goes hand in hand with the inertia aspect. But, as noted on Nabble in December, early withdrawal evaluation is really hard. I don't know how to do it properly  this is why I was _always_ against implementing it. I have only limited understanding of the mechanics of population models so far  one natural idea would be mapping the quantal response probabilities to tariff shares. 
In reply to this post by grampajohn
Of course the whole procedure was meant to include the current tariff which would then also be considered in the logit quantal response equilibrium. I like the idea of inertia and added the corresponding subutility and weight to the utility function. 
I added some more implementaitonoriented material on tariff choice to the game specs.

Administrator

Hi Chris, Thanks for the updates to the spec! Please have a look at the presentation (\papers\talks\TUDelft11ALG) I gave in Delft in Feb this year, since it includes a nice figure (maybe it could potentially replace the current one, or merge the two figures). Starting from slide 23. I also wrote down the generalized utility functions there. Please have a look, and the potential attributes. Maybe we should include the attribute table in the spec? Thoughts? Thanks, Wolf From: chris.flath [via Power TAC Developers] [mailto:[hidden email]] I added some more implementaitonoriented material on tariff choice to the game specs. If you reply to this email, your message will be added to the discussion below: http://powertacdevelopers.975333.n3.nabble.com/Tariffselectionproblemtp2833549p2842849.html To start a new topic under Power TAC Developers, email [hidden email] Disclaimer
De informatie verzonden in dit email bericht inclusief de bijlage(n)
is vertrouwelijk en is uitsluitend bestemd voor de geadresseerde van
dit bericht. Lees verder:
http://www.eur.nl/emaildisclaimer.
The information in this email message is confidential and may be legally privileged. Read more: http://www.eur.nl/english/emaildisclaimer. 
Free forum by Nabble  Edit this page 