Modelling elements

Revenues

Renewables clusters and thermal are considered equivalent for the revenue computation. No Cfd or others contracts out of the market implemented.
No reserves mechanism.
Service system are implemented. it is an amount to be distributed in proportion to the shares allocated to the technologies of the existing powerfleet.
The risk modelisation is a cVar on energy-only revenues implemented on the build_market_info_supply(). Bids is then computed on the capacity market. So it is the capacity market clearing price of the mean energy only revenue.
Hypothesis costs (investment or operational) for capacity are not accurate for renewables and dsr. Need to be filled or updated.

Batteries are not supported for investment or retirement. It may seem to work but the Antares architecture is complex : thermal cluster + link to area named z_batteries and there could be not effect.
When investing, unitcounts are added to an existing Antares cluster and timeseries are updated (otherwise, there is no effect). Then, there is no difference between older capacities and the new entrant.
No possibility to make a totally new entrant or new technology. Indeed, investment is based on NPV so revenues computed on existing capacity is compulsory.

Capacity mechanism is not delayed as it is in real life for France. Auction, and incomes from the mechanism are on the same year.
As we don't have the capacity certified computed for hydropower, it is compulsory to bypass this issue to build the supply curve with using a margin, margin = totalSupply - totalLoad at peak.

The rule implemented is based on ACER 2017: With X & Y parameters,

Trigger to reach is X(%) x priceCapInital
For N times trigger reached, newPriceCap = N x Y(€) + priceCapInitial with N stands as the round value of the mean computed on n times trigger is reached for each monte carlo year.
Then, every next simulations are modified.