APM.RTOL - APMonitor Option

Main.OptionApmRtol History

Hide minor edits - Show changes to markup

Added lines 12-18:

Python Gekko Example

(:source lang=python:) m.options.RTOL=1e-6 m.options.OTOL=1e-6 (:sourceend:)

November 26, 2019, at 01:45 PM by 136.36.211.159 -
Changed line 11 from:

RTOL is the relative inequality or equality equation tolerance for reporting a successful solution. A lower value of RTOL, such as 1e-8, will give a more precise answer but at the expense of more iterations. The default of 1e-6 is generally sufficient for most problems. However, there are times when the equation solution should be reported more precisely. Making RTOL too small may cause a bad solution to be reported because it surpasses the computer precision. RTOL and OTOL (relative tolerance for the objective function) should generally be adjusted together.

to:

RTOL is the error tolerance for reporting a successful solution where the residuals (r) of the equations (f(x)=r) are converged to nearly zero. A lower value of RTOL, such as 1e-8, will give a more precise answer but at the expense of more iterations. The default of 1e-6 is generally sufficient for most problems. However, there are times when the equation solution should be reported more precisely. Making RTOL too small may cause a bad solution to be reported because it surpasses the computer precision. RTOL and OTOL (tolerance for the objective function) should generally be adjusted together.

June 09, 2017, at 12:28 AM by 10.5.113.159 -
Changed lines 5-6 from:
to:
Added lines 12-13:

See also OTOL, SOLVER

June 01, 2017, at 07:47 PM by 45.56.3.173 -
Added lines 1-11:

(:title APM.RTOL - APMonitor Option:) (:keywords APM.RTOL, Optimization, Estimation, Option, Configure, Default, Description:) (:description Equation solution tolerance:)

 Type: Floating Point, Input
 Default Value: 1.0e-6
 Description: Equation solution tolerance

RTOL is the relative inequality or equality equation tolerance for reporting a successful solution. A lower value of RTOL, such as 1e-8, will give a more precise answer but at the expense of more iterations. The default of 1e-6 is generally sufficient for most problems. However, there are times when the equation solution should be reported more precisely. Making RTOL too small may cause a bad solution to be reported because it surpasses the computer precision. RTOL and OTOL (relative tolerance for the objective function) should generally be adjusted together.