Tuning Performance of the Network Optimizer

The default values of parameters controlling the network optimizer are generally the best choices for effective performance. However, the following sections indicate parameters that you may want to experiment with in your particular problem.

Controlling Tolerance

You control the feasibility tolerance for the network optimizer through the feasibility tolerance parameter. In the Interactive Optimizer, use the command set network tolerances feasibility.

Likewise, you control the optimality tolerance for the network optimizer through the optimality tolerance parameter. Table 6.1 and Table 6.2 summarize the default value, range, and parameter name.

 


Table 6.1    Network Tolerance Parameter: Optimality

  Optimality tolerance 
Default Value  1e-6 
Range  0.1  -  1e-11 
Callable Parameter  CPX_PARAM_NETEPOPT 
Interactive Option  network tolerances optimality 



  


Table 6.2    Network Tolerance Parameter: Feasibility

  Feasibility tolerance 
Default Value  1e-6 
Range  0.1  -  1e-11 
Callable Parameter  CPX_PARAM_NETEPRHS 
Interactive Option  network tolerances feasibility

Table 6.1 Network Tolerance Parameter: Optimality
	Optimality tolerance
Default Value	1e-6
Range	0.1 - 1e-11
Callable Parameter	`CPX_PARAM_NETEPOPT`
Interactive Option	`network tolerances optimality`

Table 6.2 Network Tolerance Parameter: Feasibility
	Feasibility tolerance
Default Value	1e-6
Range	0.1 - 1e-11
Callable Parameter	`CPX_PARAM_NETEPRHS`
Interactive Option	`network tolerances feasibility`

Selecting a Pricing Algorithm for the Network Optimizer

On the rare occasions when the network optimizer seems to take too long to find a solution, you may want to change the pricing algorithm to try to speed up computation. In the Interactive Optimizer, use the command set network pricing i, substituting a value for i to indicate which pricing algorithm to use. All the choices use variations of partial reduced-cost pricing.

Limiting Iterations in the Network Optimizer

Use the command set network iterations i, substituting a value for i, if you want to limit the number of iterations that the network optimizer performs.

Changing Sense: from Min to Max

To change a minimization problem to a maximization problem in the Interactive Optimizer, use the command change sense max and optimize again. For example, here is a transcript of a session in the Interactive Optimizer where we have already entered nexample.net and optimized it, and we now change its sense and optimize again:

CPLEX> change sense max 
Problem is a minimization problem. 
Problem is now a maximization problem. 
CPLEX> netopt 
Iteration log . . . Iteration: 0  Objective =          269.000000
Network - Optimal:  Objective =    5.0400000000e+02 
Solution time =    0.00 sec.  Iterations = 5 (0)

Because we had already solved this example once as a minimization problem, the maximization started from a feasible solution. You control whether or not the network optimizer starts from an existing solution: use the command set advance 1 to indicate in the Interactive Optimizer that you want to start from an advanced basis. This setting is the default.