Initial Solution
Two different starting strategies are available.
The first option starts with all arc flows equal to zero and
an all artificial initial basis. The second strategy gathers
the initial flows from the values shown on the worksheet.
If a basis is displayed, the arcs listed are used as the initial
basis. The algorithm starts with these values for the flows
and the basis. The advanced start option will result in fewer
iterations when only slight modifications are made to the
network model.
MIP Tolerance
When solving all integer or mixed integer programming
problems, vertices of the enumeration tree are fathomed when
the relaxed solution objective is less than (for a maximization
problem) than the incumbent solution. This tolerance makes
the fathoming test a little easier by fathoming the vertex
if its relaxed objective is within x% of the incumbent solution,
where x is the number entered in this field. The solution
may be found with fewer iterations and less time when this
value is greater than 0. When the tolerance is other than
0, however, there is no guarantee that the optimum solution
is obtained. It is a guaranteed that the solution is with
x% of the optimum. The Excel Solver has a similar tolerance
specified in the Options dialog of the Solver.
Time Limit
When solving all integer or mixed integer programming
problems the process may take a long time. The program will
stop when this time limit is reached. The user may give up
at this point or continue the optimization. The Excel Solver
has a similar time limit specified in the Options dialog of
the Solver.
Update Incumbent on the Screen
When solving all integer or mixed integer programming problems,
the algorithm may find feasible solutions during the enumeration
process. The feasible solution with the best value is called
the incumbent solution. Although intermediate steps are not
displayed on the model worksheet while the process continues,
when this checkbox is selected, the display will be changed
whenever a new incumbent is discovered.