I have added to the gallery the beginnings of a object orientated analysis of energy credits which can form both the beginnings of a software simulation and a model for understanding energy credits.

Excellent, this is something I have been looking forward to for some time.

Any thoughts on what language to use for software simulation ? (MS Visual studio BASIC.NET stuff ?)

Or perhaps there is a more specialist language that is simplier to use for the task.

Nanos wrote:



MS Visual studio BASIC.NET? curse on thou forked tonged heathen, thou speaketh the words of the devil.

anyway more productively


I can draw up a model of that in few hours in either c++ or matlab, though im not sure exactly what it is supposed to represent, it lacks come clarity

id have to say that to get a model to work an "energy credit" has to be a defined unit of measurement, what it is defined as im not certain but it has to be defined probably as either a specific quantity of energy or a specific quantity of work

also most of a model would have to be done in SI units to be a reliable and realistic model, should define as something small and simple that is not likely to change much even with increased production efficiency.


correct me if i have misinterpreted some of the model but

I think there is an inaccuracy in the apparent closed loop nature of that model

as any model involving energy has a definite sources (the Sun & Geothermal) and sinks (inefficiency and entropy)

(unless you want to model the whole universe which is closed loop but a pain to do)

looking further at the model, in general a person consumes goods rather than requires them.

personally i would view energy and a quantity or property rather than a object

i have been thinking about energy in terms of usage for other reasons (simulated Game universes)

and have been thinking about it in this way

with several interacting classes of objects


Energy Production:
Consumes: Energy
Maintenance: Goods
Produces: Energy
Requires: Work

Raw Material Production:
Consumes: energy
Maintenance: Goods
Requires: Work
Produces: Goods

Manufacturing
Consumes: Energy and Goods
Maintenance: Goods
Produces: Goods
Requires: Work

Person:
Consumes: Goods
Produces: Work
Requires: Housing, Social

Social Structure
Consumes: Energy and Goods
Maintenance: Goods
Produces: Social

Accommodation
Consumes: Energy and Goods
Maintenance: Goods
Produces: Housing

the difference between maintenance and consumption is that maintenance is a continuous requirement where as consumption is only when producing


Energy Credits only really needs to be used for a small part of the model where people turn there "allocated production capacity"
into physical Goods for their consumption

Goods are any physical consumable which usually are specified as a rate e.g units per day for each consuming or producing class


Social and work are less easily definable
work simplistically can be defined as person/hours

social is not easy to define but includes entertainment etc.

the rest of the system can use SI units


if anyone is interested i can produce a example program demonstrating this with a set of example goods

the underling models are quite complicated even for small number of goods

but for each individual end good, can produce a tree diagram

which can be used to determine its "energy value" but to be valid energy value calculations have to be done simultaneously and dynamically during simulation

but each process has to be predefined at the start of the simulation

a process being the internal conversion of goods types inside of a class
which determines its production and consumption rates.

they are defined similar to this

Consumed goods
Consumed Energy
Efficiency(ies)
Produced Goods
Produced Energy
Produced Waste
Produced Bi products

efficiencies determine additional goods and energy consumed
and additional waste and bi products produced for each unit of goods or energy produced

To finish to define an accurate model/simulation would require defining accurate process information (models) for each production class

Post edited by: fireblade, at: 2007/08/14 14:33

the approximate calculation for energy value of any specific good

is derived from

Total Energy value = SUM(Produced)+SUM(Waste)*(1/wasteeffiency)+Sum(BIproducts*(1/piprodeffiency))+SUM(Maintanance)/Time Period+WorkenergyValue

this assumes that the prduction apperatus is already in place and "paid" for

for new production apperatus have to add


(SUM(ContructGoods)+constructWorkEnergyValue)/NumberProducedoverLifetime

fireblade wrote:




I was actually thinking Java as its easy to make a graphics display.

fireblade wrote:


That’s cos its no where near finished year. I put it up to get some feed back.

fireblade wrote:


Initially I was thinking some arbitrary unit like “energy unit” just to get the sim up and running. I was then thinking that as we gather more data for the energy survey we could use real units like kwh that better matches the real situation as regards energy production and consumption in Europe.


fireblade wrote:


I don’t think we really need to model the sun as the system takes in the energy from the sun as wind, solar panels or chemically as bio fuels. At the moment I just have wind and solar but could also add bio fuels (bio-diesel, ethanol and bio-gas) under chemical energy. Losses we could model either implicit or explicit. We could, for example say production takes x amount of energy which includes loses or we could show the losses as separate. What do you think?

fireblade wrote:


Ah, some good points there.

fireblade wrote:


Yeap, got that but I need all the other uses you mentioned.

fireblade wrote:


Please do. Could we then use its as the beginnings of an energy credit sim?

fireblade wrote:


We want it as accrete as possible, like “world 3” but I think we should start with something a bit more abstract, get some understanding of it and then add to it making it more and more realistic. Sound good?

(Watches with enthusiasm.)

You know, it could be the basis for a commercial MMORPG in time, as the best way to test it once its built, is with real people..

nothing to kill so I think ppl will get board.

It wouldn't be a real simulation unless people could kill each other

Death should be factored in, and that might be one of the most appealing aspects to the game, that you only get one life..

Though I imagine it could also be more of a stratagy game, eg. you might be head of a family unit, or village elder.

There are non killing MMORPG's out there too as far as I remember.

Nanos wrote:


give me a development team and £ 1,000,000 startup cash and ill think about it


give me a few days and ill produce an example
plus a set of simple models

will most likely be in c++ or delphi, though could try and write in php if someone realy wants it





a bit more information for anyone interested

there is also the subdivision of goods and raw materials

the main difference being

Goods are discrete units

e.g
2 loaves of bread
1 bicycle
5 shirts


where as raw materials are defined by SI units (usually mass)
e.g

50 Tonnes of coal
25 tonnes of Iron
2.5 tonnes of chromium
1 tonne of nickel

energy values are not directly defined for raw materials
instead they have effectively an Energy Density in J/kg with the energy required to produce kg on that raw material

so to get there energy value you multiply the mass by the energy density

e.g 25 Mg of something with energy density of 1 MJ/Kg

1 MJ/Kg = 1000 J/Kg


so 25*10^6 * 1*10^3 = 25*10^9 J or 25 GJ for 25 Mg


Note: Mg - Mega grammes (a metric tonne)

I think it could be good if it was web based with graphics showing each part of the model and the state of the model (as graphs?). Then we could ruin it on this site.

I do like the idea of web based, perhaps Ajax ?

Nanos wrote:


Ajax, maybe depends whether brings any advantages


Ok web based is certainly possible (upto a certain level of complexity beyond which a webased front end could be done)

that would mean doing in php


this brings up some other questions though

what server will it be running on? this forum server?
what OS is it ? windows? linux?
what version of php does it have on it ? 4? 5?
what database server does it have? Mysql? Firebird? MSQQL? PostgreSQL?
how much access do we have to server to install additional libraries? (statistics and graph plotting)
How much Load can server take?

For simple example first two only matter
for more graphical third one is also important

and last becomes more important the more complex the model and more people viewing/interacting with it

and if large complex model would have to separate the two

where the model/simulation is separate from the interface and runs independently of it and is written in a compiled language

and the web interface runs separately passing in and retrieving data as required

so the simulation does not have to deal with graphics and user interfacing.

for statistics and graphics would think of using something like www.graphviz.org/ and www.aditus.nu/jpgraph/

Post edited by: fireblade, at: 2007/08/16 12:49

I would say make it either web-based or java. I think that would be more convenient for people.

Paul wrote:


I would say Use what is best for doing the job, rather than what is most convenient, though im not saying that convenience does n`t factor into that.

Also what is the intended long term direction of this ?

demonstration or accurate simulation model?

home computer or server based?

my point being that structure of implementation and best programing language(s) can vary dependent on what you are trying to achieve.

e.g for a accurate server based simulation i would think a web based (php, java or hybrid) front end and C++ based server simulation

for a home computer demonstration would be C++ or possibly java

for a server based demonstration a web based php or php/java hybrid version would probably be best

and for a accurate home computer implementation C++ i think would work best though java may also work

(well written c++ can be compiled on most OS as long as OS specific code is avoided)

these are just initial impressions might change if look at requirements more closely

I might suggest going for a server based version on the basis that a home based version would not be able to simulate accurately the most important parts of the simulation, eg. human behaviour.

That only by having human input can we be sure that our theories have some basis on being correct, else it would be too easy for a single home simulation to look like it works, but as soon as you put real people into the decision making process, you'd find out it had flaws!

A game would also attract players, who might become members, and once out of beta, you could look seriously at charging a small fee (like you would membership..) to cover your running costs.

One game aspect which appeals to gamers (I'm one myself..) is accurate simulation, less so than fancy awe inspiring graphics (Eve-online had both, but lately, they have put too much into graphics so even with a world glass PC costing several thousands, it still struggles with the pretty dust clouds in space..), which can always be bolted on at a later date, where as its hard to bolt on accurate simulation..