Given the size and complexity of the overall problem, each Monte-Carlo year is seen as a succession of weekly optimization sub-problems. All of these problems are defined at the spatial scale of the whole interconnected system. The kernel of the software is a linear solver developed by RTE and which, once fed with adequate assumptions (availability and costs of power plants, demand level, etc.) computes operating set-points for the whole system (optimal weekly unit-commitment and hydro-thermal scheduling, with an hourly resolution).

To be more accurate, the inner linear solver carries a thin outer shell that makes it possible to deal adequately enough with some of the difficult aspects of hydro-thermal optimal unit-commitment and dispatch (presence of both integer and real variables). This shell is also meant to optimize the set of individual calls to the solver, so as to adapt constantly its (hot start/presolve) strategy and thus increase the overall software’s speed.

The task of the immediately upper level is to assess and provide to the inner optimization kernel weekly hydraulic energy volumes to generate from the different reservoirs of the system, for each week of the current Monte-Carlo year. It is indeed mandatory to materialize the coupling effect that the medium and long term management of hydraulic resources introduces between the successive short term weekly optimization sub-problems. To perform its task, this module makes use of general assumptions describing the reservoir capacities and their management strategies, as well as rainfall scenarios for the current year.

The purpose of the outer layers of the tool is to feed with relevant data both the optimization kernel and the intermediate hydraulic layer. These data can be either extracted from ad hoc data banks or generated within the tool by built-in specialized modules (many practical choices are offered for mixing the two approaches)

Most of the dataset take the shape of year-long time-series (about 9000 values for a one- hour resolution): power available for one cluster of thermal plants sharing the same technical and economic characteristics, prospects of wind generation in such or such area, maximum power than can be generated from such or such reservoir, etc.

Simulation results involve all of the variables related to the system operation (generation level for each unit, flow through each tie-line), hour by hour and for all Monte-Carlo years. In addition, the tool gives an account of the CO2 emissions, as well as an assessment of the economic performance of the whole system (various estimates such as operation costs, LMP, congestion fees, etc.)