problem about the matrix multiplication

Answered

Shutian Li

March 29, 2021 00:58

• 
• 
• 

I need to code something as follows(where x is the decision variable) 

I write it as follows:(in python) 

for ss in range(cardS):

px[ss] = sum(ppp[a,i,ss]*x[a,i] for a in range(cardA) for i in range(cardS)) 

But I have an error message: TypeError: float() argument must be a string or a number, not 'MLinExpr'. 

In addition, the constraint is as follows:

Could you please help me with this error? Thanks 

 

 

0

• Official comment

  

  Simranjit Kaur

  • Gurobi Staff

  June 23, 2025 02:57 Edited

  This post is more than three years old. Some information may not be up to date. For current information, please check the Gurobi Documentation or Knowledge Base. If you need more help, please create a new post in the community forum. Or why not try our AI Gurobot?.
• 

  Jaromił Najman

  • Gurobi Staff

  March 29, 2021 09:04 Edited

  Hi,

  Please provide a minimal working example as described in Posting to the Community Forum. In particular, please provide a minimal working example defining the sets \(\mathcal{A},\mathcal{S}\), and the variables/scalars \(x_s^a, P_{is}^a, x_i^a, \bar{w}_s, \gamma\).

  Best regards,
  Jaromił

  0
• 

  Shutian Li

  March 29, 2021 14:02 Edited

  • 
  • 
  • 

  I need to code \[ \sum_{i \in S} \sum_{a \in A} P_{is}^{a}x_{i}^{a}\] 

  I wrote it as follows, it works well in a separate python file but when I use the gurobi, it does not make sense. Here is my code: 

  for s in range(cardS): 
      px[s] = sum(P[a,i,s]*x[a,i] for a in range(cardA) for i in range(cardS)) 

  The error message is TypeError: float() argument must be a string or a number, not 'MLinExpr'. 

  SetA is a set, and range(A)= cardA

  SetS is a set, and range(S)=cardS 

  Decision variable: \[x_{s}^{a}, s\in S, a\in A\]: I define it through addMVar : 

  x = gurobipy. addMVar(shape = (cardA,cardS), lb=0, vtype=GRB.CONTINUOUS)

  The parameter \[P_{i,s}^{a}, \quad i,s \in S, a \in A\] is a 3d numpy array with the shape \[A\times S\times S\]. 

  The parameter \[\bar{w}_{s}, s \in S \] is a 1d numpy array: 

  w = np.ones(shape = cardS) 

  \[\gamma\] is a scalar.  

  0
• 

  Jaromił Najman

  • Gurobi Staff

  March 29, 2021 18:32 Edited

  Hi,

  I am not sure if the Python Matrix API is the best way to go here. Please have a look at our Matrix API webinar, Matrix API example, and Matrix API Knowledge Base articles. Without the Matrix API, you can model your constraints, e.g., as

  import gurobipy 
  from gurobipy import *
  import numpy as np
  
  m = Model("test")
  A = [0,1,2,3]
  S = [0,1,2,3,4]
  P = np.random.rand(len(A),len(S),len(S))
  gamma = 42
  w = np.ones(shape = len(S))
  
  x = m.addVars(A, S, lb=0, vtype=GRB.CONTINUOUS, name="x")
  
  px = m.addConstrs((quicksum(x[a,s] for a in A) - quicksum(quicksum(gamma*P[a,i,s]*x[a,i] for a in A) for i in S) == w[s] for s in S), name="myConstr")

  For more details on the functions used, have a look at the documentation of quicksum, addConstrs, and addVars.

  Best regards,
  Jaromił

  0

Post is closed for comments.