[p2p-research] P2P Energy Demand/Supply Algorithm and Other Responses // was Re: Update: P2P Social Currency

marc fawzi marc.fawzi at gmail.com
Tue Dec 30 21:15:42 CET 2008


Emlyn:

Thanks for debating some of the finer points. It helps me explain the model
better and evolve it in the process.

<<
> A quick quibble: these energy flows aren't really peer to peer are
> they? They are peer->bank, and bank->peer, with no direct connection
> between peers. Or I might say peer->grid, grid->peer . The only real
> difference to the current situation is that we have a lot of
> grid->peer, but not much peer->grid, mostly powerstation->grid.
>
>>

Think of Common Bank as being part of the P2P client. Like you said, a
distributed server (o virtual centralized server, or whatever semantics you
use) So then what is peer and what is bank? They're all part of the whole
process.

I will add this bit too to quell fears of common bank being some sort of
goblin in the middle. It's simply the orchestration part of the distributed
process and it is implemented as part of the p2p architecture. // note to
self: need to add this. In fact, that's why it was called "peer bank" ...

<<

> 1 - Are supply and demand connected, if so, how? It looks like you can
> just draw energy from the grid without reference to supply (except if
> it all runs out I guess), as long as you can borrow enough tokens.
> Similarly it looks as though you can just add power to the grid
> without reference to whether anyone needs it, excepting for the caps,
> but how are those set?

>>

The section on caps clearly states (maybe even multiple times) that the caps
are set based on supply and demand.... where and how can I state this to be
clear on first glance?

Also, people don't need to borrow to buy energy. Every participant
(family/household/business/person) under this model needs to have their own
generation capacity. This is an arbitrary requirement now. It could be made
part of the axiomatic logic (as a starting condition) because one of the
reasons for the P2P Social Currency model is to encourage sustainably
abundant energy production from a regulated whole, not unregulated
individual.

<<

> 2 - I am a borrower of tokens, I used the tokens to buy actual power.
> Why would I pay back my debt?
>
>>

If you already have the energy production capacity, which I guess I should
add as one of the model's axiom (note to self: need to add) then you are
borrowing either to increase that energy production capacity or buy some
appreciable asset (e.g. real estate) or invest in higher production capacity
for your goods and services, etc.

You don't borrow to throw it into the sea.  And you don't borrow to lend
(because as you borrow you lose credit points for borrowing, not for seller
ranking, until you pay the loan(s) back)

So you should have the ability to pay the loan back from your excess energy
generation (conserve your own use, and sell the excess)

When you borrow your credit points for subsequent borrowing are reduced, but
your credit points for seller ranking remain //note to self: need to clarify
this under peer credits: borrowers. If you default on payments beyond the
grace period for the loan, you will lose credit points (for borrowing only,
not for seller ranking) and you won't be able to borrow again until you have
paid your loan. The obvious way out is that using your seller ranking credit
points (which you can continue to accumulate as long as you are able to
produce energy, conserve your use, and sell the rest you can continue to
make money that you use to pay back defaulted loan or use it to make more
loans and invest in new energy generation capacity or make more products and
services to sell more and then pay off the debt.. it's really up to common
sense and it's assumed that people have common sense. This model does not
prevent someone with no common sense from completely screwing up but it does
not shove them into a hole with no way out like the current model does. //note
to self: need to clarify this under peer credits: borrowers.

<<

> I lend my collaborator
> those 5 joule tokens and get 5 peer tokens.
>
>>

You mean you get peer credits that rank you higher as a seller... (no such
thing as peer tokens.. I'll work on making the semantics less
redundant/confusing)

<<

> He uses those to buy 5
> units of power which he puts into my batteries (which I count as him
> paying back the debt to me).

>>

The tokens are virtual but they don't go to your battery. They are paid to
you as virtual tokens and you do what you want with them.... you can redeem
them for equivalent in energy from common bank if you want or use them to
get goods and services, loan them to others etc

<<

> The upshot it, with very little energy input, I get endless peer
> tokens. The more efficient my storage technology, the lower the
> required input.

>>

Here is a major point I need to explain
//note to self: need to add this under Peer Credit: Lenders and Peer
Credits: Borrowers

You don't choose whom you lend to. You don't choose whom you borrow from.

You say I have 2000 tokens to lend. Common Bank (which again can be part of
the P2P client/architecture) matches you to 1 or many borrowers on first
come first served basis or some randomized fashion.

You can't run a racket when you don't know who is borrowing from you and
whom you're lending to. It happens in the background with changing user keys
rather than user names (kept by common bank in some table), so you can never
know whom lending you or borrowing from you at any given time.

<<

> 4 - The current energy trading systems cater for peak demand being
> different from average demand, by being able to have some plants whose
> production is more expensive, but which only supply when the price
> goes high enough (as during peak demand). Your model doesn't appear to
> have any similar mechanism. This relates point 1 about no apparent
> connection between supply and demand.
>
>>

There is a connection between supply and demand: dynamic caps on peer energy
production, to make sure that supply is in sync with demand... but I'm
thinking now that rather than a "cap" per se, Common Bank would issue
general requests for supply (a given quantity tied to predicted inventory
level) with ability to cancel.

//note to self: need to add this under renamed Assuring P2P Enegry Market
Equilibrium

<<

> How does capping come from supply and demand? what is the algorithm?
>
>>

//note to self: need to add the algorithm under Assuring P2P Energy Market
Equilibrium

I hadn't thought about it till now, but thanks to your question, here is how
one possible version of a KISS algorithm (there is already a patent on this,
so it's used here as an example):

This algorithm extends the Common-Bank-mediated exchange of energy by having
a longer term storage requirement (i.e. an inventory or buffer), in addition
to the exchange queue, which adds the requirement for Common Bank to be able
to issue replenishment orders (in the needed amount) rather than having a
cap per peer.

1. Take historical inventory data (based on history of supply interactions
with Common Bank) and calculate a Gaussian distribution with a Mean and
Covariance and provide an inventory level prediction at a set time in the
future.

2. Compare the inventory level prediction to a first Common Bank-set low
inventory threshold.

3. Compare the inventory level prediction to a second Common Bank-set high
inventory threshold.

4. Solicit energy replemnishment from peers if the predicted level is below
the low inventory threshold, or cancel existing requests if the predicted
level is above than the high inventory threshold. This should happen without
allowing individual peers to sign up for a disproportionate share of the
total replenihment request, and certainly not repeatedly, i.e. to avoid
building a dependency in the market on certain peers.

5. Record time dependent data related to inventory consumption and
replenishment.

6. Calculate time dependent cummulative-forecast consumption,
cummulative-forecast replenishment, cummulative actual consumption, and
cummulative actual replenishment.

7. Use the new data to update the Gaussian distribution, Mean and
Covariance, and calculate a new predicted inventory level for the next time
period.

Voila!

//note to self: need to clarify this under Anti-Dumping and Anti-Monopoly
caps and maybe change title to Assuring P2P Energy Market Equilibrium


<<

> so what is peer to peer about this process? Just the fact that small
> players can be producers as well as consumers?
>
>>

Like I said, Common Bank can be part of the P2P client... so what isn't peer
to peer about it?

I'm going to call it Peer Bank again :D after I explain that it can be built
into the P2P client as a distributed process (or virtually centralized
process)

We're just playing with words here.

Thanks again for debating the finer points.

Marc
p.s. I'll make the required changes (noted here) in version 0.80.0
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listcultures.org/pipermail/p2presearch_listcultures.org/attachments/20081230/aa9592a5/attachment-0001.html>


More information about the p2presearch mailing list