#include "ClpConfig.h"
#include <iostream>
#include <cassert>
#include <cmath>
#include <vector>
#include <string>
#include "ClpPackedMatrix.hpp"
#include "CoinMessageHandler.hpp"
#include "CoinHelperFunctions.hpp"
#include "ClpParameters.hpp"
#include "ClpObjective.hpp"
Go to the source code of this file.
Classes | |
class | ClpModel |
class | ClpDataSave |
This is a tiny class where data can be saved round calls. More... | |
Defines | |
#define | COIN_DBL_MAX DBL_MAX |
Parameter set/get methods | |
The set methods return true if the parameter was set to the given value, false otherwise. There can be various reasons for failure: the given parameter is not applicable for the solver (e.g., refactorization frequency for the volume algorithm), the parameter is not yet implemented for the solver or simply the value of the parameter is out of the range the solver accepts. If a parameter setting call returns false check the details of your solver. The get methods return true if the given parameter is applicable for the solver and is implemented. In this case the value of the parameter is returned in the second argument. Otherwise they return false. once it has been decided where solver sits this may be redone | |
#define | COIN_CBC_USING_CLP 0x01000000 |
For advanced options 1 - Don't keep changing infeasibility weight 2 - Keep nonLinearCost round solves 4 - Force outgoing variables to exact bound (primal) 8 - Safe to use dense initial factorization 16 -Just use basic variables for operation if column generation 32 -Clean up with primal before strong branching 64 -Treat problem as feasible until last minute (i.e. |
#define COIN_DBL_MAX DBL_MAX |
Definition at line 25 of file ClpModel.hpp.
#define COIN_CBC_USING_CLP 0x01000000 |
For advanced options 1 - Don't keep changing infeasibility weight 2 - Keep nonLinearCost round solves 4 - Force outgoing variables to exact bound (primal) 8 - Safe to use dense initial factorization 16 -Just use basic variables for operation if column generation 32 -Clean up with primal before strong branching 64 -Treat problem as feasible until last minute (i.e.
minimize infeasibilities) 128 - Switch off all matrix sanity checks 256 - No row copy 512 - If not in values pass, solution guaranteed, skip as much as possible 1024 - In branch and bound 2048 - Don't bother to re-factorize if < 20 iterations 4096 - Skip some optimality checks 8192 - Do Primal when cleaning up primal 16384 - In fast dual (so we can switch off things) 32768 - called from Osi 65536 - keep arrays around as much as possible (also use maximumR/C) 131072 - scale factor arrays have inverse values at end 262144 - extra copy of scaled matrix NOTE - many applications can call Clp but there may be some short cuts which are taken which are not guaranteed safe from all applications. Vetted applications will have a bit set and the code may test this At present I expect a few such applications - if too many I will have to re-think. It is up to application owner to change the code if she/he needs these short cuts. I will not debug unless in Coin repository. See COIN_CLP_VETTED comments. 0x01000000 is Cbc (and in branch and bound) 0x02000000 is in a different branch and bound
Definition at line 852 of file ClpModel.hpp.