Soufyan Zayou

1. Community
2. Algorithms
3. Why is Gurobi is not fully exploiting the sparsity and structure of the sparse QP problem?

Hi, No, it was expected that the computational complexity would scale linearly with the prediction horizon  N, i.e. O(N) for a sparse QP problem, so the slope of the line is linear in logartimic sc...

• View comment
• Soufyan Zayou
• March 24, 2020 02:19
• Edited
• 0 votes

1. Community
2. Algorithms
3. Barrier algorithm; what kind of matrices can it handle the best (sparse or dense, large or small problems)?

Hi, As compared to simplex, where the method would generally require lots of relatively inexpensive iterations, the barrier will require much fewer iteration but each one of those is much more expe...

• View comment
• Soufyan Zayou
• March 24, 2020 02:18
• Edited
• 0 votes

1. Community
2. Algorithms
3. Model contains large quadratic objective coefficient range.

Hi, Thanks for your response. Still one general question about Gurobi (you maybe know from experience). I have some general questions about the barrier algoritm that Gurobi uses to solve a QP. For ...

• View comment
• Soufyan Zayou
• March 11, 2020 19:28
• Edited
• 0 votes

1. Community
2. Algorithms
3. Model contains large quadratic objective coefficient range.

Hi,  I have re-scaled the model, but the numerical issues are still present.  Do you have any other options I could try? What did you by the way mean with: border-line with the round-off error in d...

• View comment
• Soufyan Zayou
• March 07, 2020 18:24
• Edited
• 0 votes

1. Community
2. Algorithms
3. Model contains large quadratic objective coefficient range.

Part 2: In more detail, after some substitution the quadritic part of the optimization problem looks like:  with Z a banded null-space matrix, Omega a diagonal matrix with transpose(H)*Q*H  and R ...

• View comment
• Soufyan Zayou
• February 11, 2020 00:26
• Edited
• 0 votes

1. Community
2. Algorithms
3. Model contains large quadratic objective coefficient range.

Hi, Thanks for the clarification. I forgot to give you more information. Part 1: The optimization problem I am solving could be seen below: with zi|k = H*xi|k, ui|k the predicted inputs, xi|k the ...

• View comment
• Soufyan Zayou
• February 10, 2020 20:02
• Edited
• 0 votes

1. Community
2. Algorithms
3. Model contains large quadratic objective coefficient range.

Hello Yuriry, Does the QObjective range specify the range of the elements in the Hessian? Is is somewhat the same as the condition number of the Hessian? The right-hand side of the inequality const...

• View comment
• Soufyan Zayou
• February 10, 2020 14:52
• 0 votes

1. Community
2. Algorithms
3. Tolerance and constraint parameters

Hello Yuriy, Thanks for your reply.  I don't really understand the last sentence of your answer: ''indeed, linear algebra --a driving engine, so to speak-- behind the barrier is often more efficien...

• View comment
• Soufyan Zayou
• February 10, 2020 11:19
• 0 votes

1. Community
2. Programming
3. Model contains large quadratic objective coefficient range.

Hi Silke, Thanks for your response. I have still a couple of questions based on your answer. Which numerical properties may change when you use scaling? Could you give examples? Furthermore, you sa...

• View comment
• Soufyan Zayou
• February 04, 2020 13:39
• 0 votes

1. Community
2. Programming
3. Solving QP problem using Gurobi in Matlab wrong optimization results

This is the output file, where it is stated that is a suboptimal solution. Does anyone know what is wrong? Gurobi Optimizer version 9.0.0 build v9.0.0rc2 (win64)Optimize a model with 1200 rows, 600...

• View comment
• Soufyan Zayou
• January 16, 2020 15:16
• 0 votes