When I try to duplicate results obtained for a simple linear program using lpSolve or gurobi, I run into issues. In particular, the function solvecop throws a message "Certificate of dual infeasibility found" (the solver employed is cccp).
Upon closer investigation, I realized that the issue comes from line 17 in the function simplify.R
In this line, the object a is kind of normalized and re-scaled. When I uncomment this line, I recover the same results as when using lpSolve/gurobi. I also recover the results when using the cccp function directly.
What is the reason for this normalization/re-scaling in simplify.R?
When I try to duplicate results obtained for a simple linear program using lpSolve or gurobi, I run into issues. In particular, the function solvecop throws a message "Certificate of dual infeasibility found" (the solver employed is cccp). Upon closer investigation, I realized that the issue comes from line 17 in the function simplify.R In this line, the object a is kind of normalized and re-scaled. When I uncomment this line, I recover the same results as when using lpSolve/gurobi. I also recover the results when using the cccp function directly. What is the reason for this normalization/re-scaling in simplify.R?