|
|
|
mipSolve
Solves mixed-integer linear programs (MIP), with dense or sparse Matlab
matrices, using a branch-and-bound
algorithm by Nemhauser and Wolsey: Integer Programming, chap 8.2, 1989.
Main features
-
The branch-and-bound
implementation has
three types of tree searching: depth
first searching; breadth first
searching, and depth first until an integer
value solution is found, then breadth searching.
-
Priority based variable selection.
The user gives a
weight for each variable to be used in the variable selection phase.
-
The dual LP solver
that solves relaxed subproblems
is selectable.
Using MINOS in Tomlab /MINOS gives rapid
solution using MEX-file interfaces.
-
A simple knapsack heuristic is implemented, speeding up the solution
of knapsack problems.
-
The user may give an upper bound on the integer value wanted.
Makes it possible to cut branches and avoid node computations.
|
|