Uncertainty(name::String, distribution::Distribution{Univariate, Continuous})

Keyword struct for defining input uncertainties.

• The name argument can be any REeopt input
• The distribution argument can be any of the type Distribution{Univariate, Continuous} from Distributions.jl.

Example

u1 = Uncertainty(
    name="Financial.elec_cost_escalation_pct",
    distribution=Uniform(-0.01, 0.03)    
)

Distribution types

You can view all of the possible distributions with:

subtypes(Distribution{Univariate, Continuous})

Custom sample sets

For the special case of ElectricLoad.loads_kw we allow one to provide a sample set of load profiles. The sample set is sampled uniformly. To use this capability define an Uncertainty as follows:

u4 = Uncertainty(
    "ElectricLoad.loads_kw",
    [
        ones(8760) .* 8.0,  # replace these vectors with your load profile samples
        ones(8760) .* 9.0,
        ones(8760) .* 10.0,
    ]
)

Scenarios(
    base_scenario::AbstractDict, 
    uncertainties::AbstractVector{Uncertainty},
    metrics::AbstractVector{String};
    Nscenarios::Int=100,
    Nparallel::Int=1,
    ngens::Int=100
)

Args

• base_scenario::AbstractDict a Dict` that defines the base REopt Scenario (see REopt docs for more)
• uncertainties::AbstractVector{Uncertainty} vector of Uncertaintys
• metrics::AbstractVector{String} vector of output metrics (TODO use these for analysis tools)

Keyword Args

• Nscenarios::Int number of scenarios to sample from Latin Hypercube
• Nparallel::Int number of scenarios to run in parallel (when using Run Threaded Scenarios)
• ngens::Int hyperparameter of LatinHypercubeSampling.LHCoptim, used to create Latin Hypercube of samples

Scenarios(
    fp::String, 
    uncertainties::AbstractVector{Uncertainty},
    metrics::AbstractVector{String};
    Nscenarios::Int=100,
    Nparallel::Int=1,
    ngens::Int=100
)

Args

• fp::String path to a JSON file that defines the base REopt Scenario (see REopt docs for more)
• uncertainties::AbstractVector{Uncertainty} vector of Uncertaintys
• metrics::AbstractVector{String} vector of output metrics (TODO use these for analysis tools)

Keyword Args

• Nscenarios::Int number of scenarios to sample from Latin Hypercube
• Nparallel::Int number of scenarios to run in parallel (when using Run Threaded Scenarios)
• ngens::Int hyperparameter of LatinHypercubeSampling.LHCoptim, used to create Latin Hypercube of samples

run_serial_scenarios(s::Scenarios, optimizer; remove_series=false)

Run the REopt scenarios one at a time in for loop and return a vector of dictionaries for results.

If remove_series is true then all time series results are removed from the dictionaries, which will keep memory use lower and make the results compatible with rectangular data stores.

run_threaded_scenarios(s::Scenarios, optimizer; remove_series=false)

Run batches of REopt scenarios in parallel according to Scenarios.Nparallel and return a vector of dictionaries for results.

If remove_series is true then all time series results are removed from the dictionaries, which will keep memory use lower and make the results compatible with rectangular data stores.

dicts_to_dataframe(ds::AbstractVector)

Put all of the Dicts in ds into a DataFrame. Column names will match keys.

save_dataframe_as_csv(df::DataFrame, filename::String)

Needs no explanation.

save_opt_and_outage_sim_results_as_json(file_name::String, results::Vector{<:Dict}, outage_sim_results::Vector{<:Dict})

Combine optimization and outage simulator results and save them to JSON using file_name.