Added computation of cumulative minimum capacity retirements in multistage

[filippopecci]

Fixed computation of cumulative minimum retirements in multistage code. Previously, minimum capacity retirement parameters were not added up and this resulted in the minimum retirement constraints enforcing wrong lower bounds ( #514 )

[cfe316]

To make the code easier to read, consider adding a simple one-line function like

dfgen(stage) = inputs_d[stage]["dfGen"]

and also

dfvrestore(stage) = inputs_d[stage]["dfVRE_STOR"]

This also has the advantage that it makes the interface smaller. If you ever want to change "dfVRE_STOR" to another key it's easier since it's only in one place.

[cfe316]

Could also do something like

function xxx(a,b,c):
    if !isempty(a)
        inputs[...][!, Symbol(b)] = sum( .... Symbol(c) for p in 1:t)
    end
end
     
table = [("VS_DC", "Min_Cum_Retired_Cap_Inverter_MW", "Min_Retired_Cap_Inverter_MW"),
              (..., ..., ...),
              ]
for row in table:
    a,b,c = row
    xxx(a,b,c)
end

[lbonaldo]

Two more suggestions to improve readability in addition to @cfe316 comments:

• You could use a Symbol to select a column (e.g.,:Min_Cum_Retired_Energy_Cap_MW) instead of converting a String into a Symbol every time, (e.g., Symbol("Min_Cum_Retired_Energy_Cap_MW"))

• The usual convention for mutating functions (e.g., myfun!(t::Int,inputs_d::Dict) is that the argument that is modified is placed first (Ref.: https://docs.julialang.org/en/v1/manual/style-guide/#Write-functions-with-argument-ordering-similar-to-Julia-Base)

Suggested change

	function compute_min_cumulative_retirements!(t::Int,inputs_d::Dict)
	function compute_min_cumulative_retirements!(inputs_d::Dict, t::Int)

[lbonaldo]

To improve performance, If t can be very large, we could think about a way to avoid redoing the sum over all the keys 1:t of the dictionary, since the sum over 1:t-1 is stored at inputs_d[t-1]["dfGen"][!,:Min_Cum_Retired_Cap_MW]. Am I correct?

[lbonaldo]

Is this a typo?

Suggested change

	if !isempty(inputs["VS_DC"])
	if !isempty(inputs_d[1]["VS_DC"])

[filippopecci]

Thank you @cfe316 and @lbonaldo for your suggestions. I have implemented all of them, have a look and let me know.

[cfe316]

This should be under a ### Fixed section of the changelog. If we don't have one, add it below ### Added. See https://keepachangelog.com/en/1.1.0/ .

[cfe316]

Is the first mytab a subset of the second one? Could this be converted to

mytab = ...

if VRESTORE
   add new entries to my_tab
end

Then they'd only need to be listed once.

[cfe316]

It looks like the second line here is the sum of the first line plus the term inputs_d[t-1][dfGen_Name][!,CumRetCap]

Could this be reformulated to

A = B
if t==1
   A += C
end

[cfe316]

Thanks for the fix!