- Timestamp:
- Dec 5, 2009 1:34:51 PM (11 years ago)
- File:
-
- 1 edited
Legend:
- Unmodified
- Added
- Removed
-
branches/sandbox/Cbc/src/CbcNode.cpp
r1316 r1366 3528 3528 double objMax = -1.0e50; 3529 3529 bool needResolve = false; 3530 /* 3531 Now calculate the cost forcing the variable up and down. 3532 */ 3530 3533 int iDo; 3531 3534 for (iDo = 0; iDo < numberToDo; iDo++) { … … 3539 3542 int iColumn = dynamicObject->columnNumber(); 3540 3543 int preferredWay; 3544 /* 3545 Update the information held in the object. 3546 */ 3541 3547 object->infeasibility(&usefulInfo, preferredWay); 3542 3548 double value = currentSolution[iColumn]; … … 3980 3986 the up arm. But see OsiBranchingObject::way() 3981 3987 Use nodeInfo--.numberBranchesLeft_ to see how active 3988 3989 Except that there is no OsiBranchingObject::way(), and this'll fail in any 3990 event because we have various OsiXXXBranchingObjects which aren't descended 3991 from CbcBranchingObjects. I think branchIndex() is the appropriate 3992 equivalent, but could be wrong. (lh, 061220) 3993 3994 071212: I'm finally getting back to cbc-generic and rescuing a lot of my 3995 annotation from branches/devel (which was killed in summer). I'm going to 3996 put back an assert(obj) just to see what happens. It's still present as of 3997 the most recent change to CbcNode (r833). 3998 3999 080104: Yep, we can arrive here with an OsiBranchingObject. Removed the 4000 assert, it's served its purpose. 4001 4002 080226: John finally noticed this problem and added a way() method to the 4003 OsiBranchingObject hierarchy. Removing my workaround. 4004 3982 4005 */ 3983 4006 int
Note: See TracChangeset
for help on using the changeset viewer.