In this example we demonstrate the use of Box constraints in xSolverPro to restrict the
range of variables in the solution set. Consider the hatfb problem from CUTE. The function
to minimize is defined as:
x[0]=p[0]-1
x[1]=p[0]-sqrt(p[1])
x[2]=p[1]-sqrt(p[2])
x[1]=p[2]-sqrt(p[3])
This function has a minimum at {1, 1, 1, 1} when unconstrained, but what if we want to make
p[1] have an upper bound of 0.8 and set a lower bound of 0 for p0..p3?
We start by defining a spreadsheet with our functions as in figure 1 and setting the
variables and residuals as explained in the example using xSolverLite to fit a sine wave with
a polynomial. Open the Setup Screen.
range of variables in the solution set. Consider the hatfb problem from CUTE. The function
to minimize is defined as:
x[0]=p[0]-1
x[1]=p[0]-sqrt(p[1])
x[2]=p[1]-sqrt(p[2])
x[1]=p[2]-sqrt(p[3])
This function has a minimum at {1, 1, 1, 1} when unconstrained, but what if we want to make
p[1] have an upper bound of 0.8 and set a lower bound of 0 for p0..p3?
We start by defining a spreadsheet with our functions as in figure 1 and setting the
variables and residuals as explained in the example using xSolverLite to fit a sine wave with
a polynomial. Open the Setup Screen.
Figure 1
When the setup screen opens, press the 'Solver >>' button to expose the options and
additional buttons so that it appears as in Figure 2. Notice the 'Set Constraints' button in
the lower left. Press it to bring up the 'Set Constraints' screen as in Figure 3.
additional buttons so that it appears as in Figure 2. Notice the 'Set Constraints' button in
the lower left. Press it to bring up the 'Set Constraints' screen as in Figure 3.
Figure 2
The 'Set Solver Constraints' screen should be open as in Figure 3. Press the
'Box' radio button under 'Constraint Type' .
'Box' radio button under 'Constraint Type' .
Figure 3