Help! Dynamic constraint definition for power system MIP optimization in Gurobi

ユーザーの入力を待っています。

Elliott Fix

2022年04月23日 02:39

To whom it may concern at Gurobi Support,

I hope you are doing well.

I have below some Python code that I would like to convert to a Gurobi constraint, however, I am running into issues. I am trying to solve an MIP Unit Commitment problem .

The code that I'd like to convert to a Gurobi constraint is:

for t in time_periods:
    UT = min(gen['time_up_minimum'],data['time_periods'])
    if t >= UT:
        m.uptime[g,t] = sum(m.vg[g,t] for t in range(t-UT+1, t+1)) <= m.ug[g,t]

UT is a list that contains data of the uptime of each generator each hour.

 

My attempt at making this Python loop a Gurobi constraint is below. Note that I use the variable 'p' as the equivalent iterator 't'

min_uptime = model.addConstrs(gp.quicksum(nstart[g, p] for p in range(p-UT+1,p+1)) <= ngen[g, periods]
                            for g in range(ngens) for periods in range(nperiods))

The error that results when I run my code is

File C:/Users/... test.py:60 in <genexpr>
    min_uptime = model.addConstrs(gp.quicksum(nstart[g, p] for p in range(p-UT+1,p+1)) <= ngen[g, periods]

NameError: name 'p' is not defined

I know that the issue stems from the part:

... for p in range(p-UT+1,p+1)) ...

I believe the issue is that this is a dynamic constraint, meaning the constraint is not constant; it changes every time period, 'p'. I cannot think of a way to include a way to iterate through p inside the range like the Python code iterates with 't'.

What do you think?

Thanks in advance!!!

 

Elliott Fix

 

0

• 

  Jaromił Najman

  • Gurobi Staff

  2022年04月25日 08:17

  Hi Elliott,

  In your code that you would like to convert, it would be best to not misuse \(\texttt{t}\) as iterator and fixed value. I think that the following should also work for you

  for T in time_periods:
      UT = min(gen['time_up_minimum'],data['time_periods'])
      if T >= UT:
          m.uptime[g,t] = sum(m.vg[g,t] for t in range(T-UT+1, T+1)) <= m.ug[g,T]

  You can then convert the above to Gurobi code by looping over time_periods and adjusting \(\texttt{p}\) to \(\texttt{T}\) in the \(\texttt{range}\) function.

  Best regards, 
  Jaromił

  0

サインインしてコメントを残してください。