KNITRO - Revision history

Ango: /* Description of Inputs */ FTOL, XTOL

2019-03-06T13:18:35Z

Description of Inputs: FTOL, XTOL

← Older revision		Revision as of 08:18, 6 March 2019
Line 259:		Line 259:
	The Knitro built-in functions are based on standard netlib routines (www.netlib.org). The Intel MKL functions are written especially for x86 processor architectures. On a machine running an Intel processor (e.g., Pentium 4), testing indicates that the MKL functions can reduce the CPU time in BLAS/LAPACK operations by 20-30%.		The Knitro built-in functions are based on standard netlib routines (www.netlib.org). The Intel MKL functions are written especially for x86 processor architectures. On a machine running an Intel processor (e.g., Pentium 4), testing indicates that the MKL functions can reduce the CPU time in BLAS/LAPACK operations by 20-30%.

	If your machine uses security enhanced Linux (SELinux), you may see errors when loading the Intel MKL.		If your machine uses security enhanced Linux (SELinux), you may see errors when loading the Intel MKL.
	\|-valign="top"		\|-valign="top"
	\|''DEBUG''\|\|Controls the level of debugging output. Debugging output can slow execution of Knitro and should not be used in a production setting. All debugging output is suppressed if option "PriLevOpt" equals zero. Default value: 0.		\|''DEBUG''\|\|Controls the level of debugging output. Debugging output can slow execution of Knitro and should not be used in a production setting. All debugging output is suppressed if option "PriLevOpt" equals zero. Default value: 0.
Line 274:		Line 274:
	\|-valign="top"		\|-valign="top"
	\|''FEASTOL_ABS''\|\|Specifies the final absolute stopping tolerance for the feasibility error. Smaller values of FEASTOLABS result in a higher degree of accuracy in the solution with respect to feasibility. Default 0. '''NOTE''': For more information on the stopping test used in TOMLAB /KNITRO see [http://tomwiki.com/KNITRO#Termination_Test_and_Optimality].		\|''FEASTOL_ABS''\|\|Specifies the final absolute stopping tolerance for the feasibility error. Smaller values of FEASTOLABS result in a higher degree of accuracy in the solution with respect to feasibility. Default 0. '''NOTE''': For more information on the stopping test used in TOMLAB /KNITRO see [http://tomwiki.com/KNITRO#Termination_Test_and_Optimality].
			\|-valign="tp"
			\|''FTOL''\|\|The optimization process will terminate if the relative change in the objective function is less than FTOL for FTOL_ITERS consecutive feasible iterations. Default 1''e - ''15.
			\|-valign="tp"
			\|''FTOL_ITERS''\|\|See FTOL for explanation. Default: 5.

	\|-valign="tp"		\|-valign="tp"
	\|''GRADOPT''\|\|		\|''GRADOPT''\|\|
Line 558:		Line 563:
	\|-valign="top"		\|-valign="top"
	\|''XTOL''\|\|The optimization will terminate when the relative change in the solution estimate is less than XTOL. If using an interior-point algorithm and the barrier parameter is still large, TOMLAB /KNITRO will first try decreasing the barrier parameter before terminating. Default 1''.''0''e - ''15.		\|''XTOL''\|\|The optimization will terminate when the relative change in the solution estimate is less than XTOL. If using an interior-point algorithm and the barrier parameter is still large, TOMLAB /KNITRO will first try decreasing the barrier parameter before terminating. Default 1''.''0''e - ''15.
			\|-valign="top"
			\|''XTOL_ITERS''\|\|The optimization will terminate if the relative change in the solution estimate is less than XTOL for XTOL_ITERS consecutive iterations. If set to 0, KNITRO chooses this value based on the algorithm and context. Currently KNITRO sets this value to 3 unless the MISQP algorithm is being used, in which case the value is set to 1 by default."
	\|}		\|}

Ango: /* Active */

2016-11-21T11:20:27Z

Active

← Older revision		Revision as of 06:20, 21 November 2016
Line 837:		Line 837:
	===Active===		===Active===

	Knitro 4.0 ~~introduces~~ a new active-set Sequential Linear-Quadratic Programing (SLQP) optimizer. This ~~opti- mizer~~ is particular advantageous when "warm starting" (i.e., when the user can provide a good initial solution estimate, for example, when solving a sequence of closely related problems). This algorithm is also the preferred algorithm for detecting infeasible problems quickly. The Active algorithm can be chosen by setting ALG=3. It can use any of the Hessian options.		Knitro 4.0 introduced a new active-set Sequential Linear-Quadratic Programing (SLQP) optimizer. This optimizer is particular advantageous when "warm starting" (i.e., when the user can provide a good initial solution estimate, for example, when solving a sequence of closely related problems). This algorithm is also the preferred algorithm for detecting infeasible problems quickly. The Active algorithm can be chosen by setting ALG=3. It can use any of the Hessian options.

			The fourth available method, an SQP algorithm ALG=4) in Knitro is an active-set method that solves a sequence of quadratic programming subproblems to solve the problem. This method is primarily designed for small to medium scale problems with expensive function evaluations – for example, problems where the function evaluations involve performing expensive black-box simulations and/or derivatives are computed via finite-differencing.

	==Other special features==		==Other special features==

Ango: /* Gradient Checks */

2013-07-31T16:09:01Z

Gradient Checks

← Older revision		Revision as of 11:09, 31 July 2013
Line 857:		Line 857:
	====Gradient Checks====		====Gradient Checks====

	If the user is supplying a routine for computing exact gradients, but would like to compare these gradients with finite-difference gradient approximations as an error check this can easily be done in Knitro. ~~To perform a gradient check with ''forward ''finite~~-~~differences set GRADOPT~~=~~4, and~~ to ~~perform a~~ gradient ~~check with ''centered ''finite-differences set GRADOPT=5~~.		If the user is supplying a routine for computing exact gradients, but would like to compare these gradients with finite-difference gradient approximations as an error check this can easily be done in Knitro. Set

			Prob.KNITRO.checkderiv = 1

			and optionally, either or both of

			Prob.KNITRO.tol_rel = 1e-6
			Prob.KNITRO.tol_abs = 1e-6

			(default values are shown here) to control how large/small discrepancies will be shown. Also enable the printfile feature to see the gradient checking output which is printed before the main solve starts.

	===Second derivative options===		===Second derivative options===

Ango: /* Description of Inputs */

2012-12-21T11:01:14Z

Description of Inputs

← Older revision		Revision as of 06:01, 21 December 2012
Line 193:		Line 193:
	3: Implement both options 1 and 2 above.		3: Implement both options 1 and 2 above.

	If "~~bar feasible~~=1" or "~~bar feasible~~=3", this will activate the feasible version of Knitro. The feasible version of Knitro will force iterates to strictly satisfy inequalities, but does not require satisfaction of equality constraints at intermediate iterates (see section 8.3). This option and the "honorbnds" option may be useful in applications where functions are undefined outside the region defined by inequalities. The initial point must satisfy inequalities to a sufficient degree; if not, Knitro may generate infeasible iterates and does not switch to the feasible version until a sufficiently feasible point is found. Sufficient satisfaction occurs at a point x if it is true for all inequalities that, ''cl ''+ ''tol = c''(''x'') ''= cu - tol ''(2) (for ''cl ''= ''cu''). The constant ''tol ''is determined by the option ~~bar feasmodetol~~.		If "bar_feasible=1" or "bar_feasible=3", this will activate the feasible version of Knitro. The feasible version of Knitro will force iterates to strictly satisfy inequalities, but does not require satisfaction of equality constraints at intermediate iterates (see section 8.3). This option and the "honorbnds" option may be useful in applications where functions are undefined outside the region defined by inequalities. The initial point must satisfy inequalities to a sufficient degree; if not, Knitro may generate infeasible iterates and does not switch to the feasible version until a sufficiently feasible point is found. Sufficient satisfaction occurs at a point x if it is true for all inequalities that, ''cl ''+ ''tol = c''(''x'') ''= cu - tol ''(2) (for ''cl ''= ''cu''). The constant ''tol ''is determined by the option bar_feasmodetol.

	If "~~bar feasible~~=2" or "~~bar feasible~~=3", Knitro will place special emphasis on first trying to get feasible before trying to optimize.		If "bar_feasible=2" or "bar_feasible=3", Knitro will place special emphasis on first trying to get feasible before trying to optimize.
	\|-valign="top"		\|-valign="top"
	\|''BAR_FEASMODETOL''\|\|Specifies the tolerance in equation 2 that determines whether Knitro will force subsequent iterates to remain feasible. The tolerance applies to all inequality constraints in the problem. This option only has an effect if option "~~bar feasible~~=1" or "~~bar feasible~~=3". Default 1''e - ''4.		\|''BAR_FEASMODETOL''\|\|Specifies the tolerance in equation 2 that determines whether Knitro will force subsequent iterates to remain feasible. The tolerance applies to all inequality constraints in the problem. This option only has an effect if option "bar_feasible=1" or "bar_feasible=3". Default 1''e - ''4.
	\|-valign="top"		\|-valign="top"
	\|''BAR_INITMU''\|\|Specifies the initial barrier parameter value for the interior-point algorithms. This option has no effect on the Active Set algorithm. Default 1''.''0''e - ''1.		\|''BAR_INITMU''\|\|Specifies the initial barrier parameter value for the interior-point algorithms. This option has no effect on the Active Set algorithm. Default 1''.''0''e - ''1.

Ango: /* Introduction */ Hyphenations eliminated.

2012-12-21T10:58:13Z

Introduction: Hyphenations eliminated.

← Older revision		Revision as of 05:58, 21 December 2012
Line 5:		Line 5:
	Welcome to the TOMLAB /KNITRO User's Guide. TOMLAB /KNITRO includes the KNITRO (MI)NLP solver from Ziena Optimization and an interface to The MathWorks' MATLAB.		Welcome to the TOMLAB /KNITRO User's Guide. TOMLAB /KNITRO includes the KNITRO (MI)NLP solver from Ziena Optimization and an interface to The MathWorks' MATLAB.

	Knitro implements both state-of-the-art interior-point and active-set methods for solving nonlinear optimization problems. In the interior method (also known as a barrier method), the nonlinear programming ~~prob- lem~~ is replaced by a series of barrier sub-problems controlled by a barrier parameter ''µ ''(initial value set using ''Prob.KNITRO.options.BAR_INITMU''). The algorithm uses trust regions and a merit function to promote ~~con- vergence~~. The algorithm performs one or more minimization steps on each barrier problem, then decreases the barrier parameter, and repeats the process until the original problem has been solved to the desired accuracy.		Knitro implements both state-of-the-art interior-point and active-set methods for solving nonlinear optimization problems. In the interior method (also known as a barrier method), the nonlinear programming problem is replaced by a series of barrier sub-problems controlled by a barrier parameter ''µ ''(initial value set using ''Prob.KNITRO.options.BAR_INITMU''). The algorithm uses trust regions and a merit function to promote convergence. The algorithm performs one or more minimization steps on each barrier problem, then decreases the barrier parameter, and repeats the process until the original problem has been solved to the desired accuracy.

	Knitro provides two procedures for computing the steps within the interior point approach. In the version known as Interior/CG each step is computed using a projected conjugate gradient iteration. This approach differs from most interior methods proposed in the literature in that it does not compute each step by solving a linear system involving the KKT (or primal-dual) matrix. Instead, it factors a projection matrix, and uses the conjugate gradient method, to approximately minimize a quadratic model of the barrier problem.		Knitro provides two procedures for computing the steps within the interior point approach. In the version known as Interior/CG each step is computed using a projected conjugate gradient iteration. This approach differs from most interior methods proposed in the literature in that it does not compute each step by solving a linear system involving the KKT (or primal-dual) matrix. Instead, it factors a projection matrix, and uses the conjugate gradient method, to approximately minimize a quadratic model of the barrier problem.
Line 11:		Line 11:
	The second procedure for computing the steps, which we call Interior/Direct, always attempts to compute a new iterate by solving the primal-dual KKT matrix using direct linear algebra. In the case when this step cannot be guaranteed to be of good quality, or if negative curvature is detected, then the new iterate is computed by the Interior/CG procedure.		The second procedure for computing the steps, which we call Interior/Direct, always attempts to compute a new iterate by solving the primal-dual KKT matrix using direct linear algebra. In the case when this step cannot be guaranteed to be of good quality, or if negative curvature is detected, then the new iterate is computed by the Interior/CG procedure.

	Knitro also implements an active-set sequential linear-quadratic programming (SLQP) algorithm which we call Active. This method is similar in nature to a sequential quadratic programming method but uses linear ~~program- ming~~ sub-problems to estimate the active-set at each iteration. This active-set code may be preferable when a good initial point can be provided, for example, when solving a sequence of related problems.		Knitro also implements an active-set sequential linear-quadratic programming (SLQP) algorithm which we call Active. This method is similar in nature to a sequential quadratic programming method but uses linear programming sub-problems to estimate the active-set at each iteration. This active-set code may be preferable when a good initial point can be provided, for example, when solving a sequence of related problems.

	Knitro has support for mixed-integer programming, either by a standard branch and bound method or a hybrid Quesada-Grossman method.		Knitro has support for mixed-integer programming, either by a standard branch and bound method or a hybrid Quesada-Grossman method.

Ango: /* Description of Inputs */ text fixes

2012-12-20T09:48:18Z

Description of Inputs: text fixes

Show changes

Ango: /* Description of Inputs */ Several fixes and improved links

2012-12-20T09:20:11Z

Description of Inputs: Several fixes and improved links

Ango: /* Description of Inputs */ Correct one occurrence of non-existing parameter "feasible"

2012-12-17T12:55:17Z

Description of Inputs: Correct one occurrence of non-existing parameter "feasible"

← Older revision		Revision as of 07:55, 17 December 2012
Line 201:		Line 201:
	\|''BAR_INITMU''\|\|Specifies the initial barrier parameter value for the interior-point algorithms. This option has no effect on the Active Set algorithm. Default 1''.''0''e - ''1.		\|''BAR_INITMU''\|\|Specifies the initial barrier parameter value for the interior-point algorithms. This option has no effect on the Active Set algorithm. Default 1''.''0''e - ''1.
	\|-valign="top"		\|-valign="top"
	\|''BAR_INITPT''\|\|Indicates whether an initial point strategy is used with barrier algorithms. If the "honorbnds" or "~~feasible~~" options are enabled, then a request to shift may be ignored. This option has no effect on the Active Set algorithm. Default 0.		\|''BAR_INITPT''\|\|Indicates whether an initial point strategy is used with barrier algorithms. If the "honorbnds" or "bar_feasible" options are enabled, then a request to shift may be ignored. This option has no effect on the Active Set algorithm. Default 0.

	0 (auto): Let Knitro automatically choose the strategy.		0 (auto): Let Knitro automatically choose the strategy.

Elias at 14:58, 17 December 2011

2011-12-17T14:58:36Z

← Older revision		Revision as of 09:58, 17 December 2011
Line 91:		Line 91:
	<math>		<math>
	\begin{array}{ccccccl}		\begin{array}{ccccccl}
	\min\limits_{x} & ~~\multicolumn{5}{l}{~~f(x)} & \\		\min\limits_{x} & f(x) & \\
	\\		\\
	s/t & x_L & \leq & x & \leq & x_U & \\		s/t & x_L & \leq & x & \leq & x_U & \\
Line 706:		Line 706:
	<math>		<math>
	\begin{array}{ccc}		\begin{array}{ccc}
	\~~minimize~~{x} && f(x) \\		\min{x} && f(x) \\
	\mbox{subject to} && h(x) = 0 \\		\mbox{subject to} && h(x) = 0 \\
	&& g(x) \leq 0. \\		&& g(x) \leq 0. \\
Line 716:		Line 716:
	<math>		<math>
	\begin{array}{l}		\begin{array}{l}
	\nabla_x ~~{\cal~~ L}(x,\lambda) = \nabla f(x) +		\nabla_x L (x,\lambda)= \nabla f(x) + \lambda_i\nabla h_i(x) +
	~~\sum_{i \in {\cal E}}~~\lambda_i\nabla h_i(x) +		\sum_{i \in {I}}\lambda_i\nabla g_i(x)=0\\
	\sum_{i \in {~~\cal~~ I}}\lambda_i\nabla g_i(x)=0\\		\lambda_i g_i(x) = 0, \quad i \in I\\
	\lambda_i g_i(x) = 0, \quad i \in I}\\		h_i(x) = 0, \quad i \in E\\
	h_i(x) = 0, \quad i \in ~~{\cal_E}~~\\		g_i(x) \le 0, \quad i \in I\\
	g_i(x) \le 0, \quad i \in ~~{\cal_I}~~\\		\lambda_i \ge 0, \quad i \in I
	\lambda_i \ge 0, \quad i \in {\~~cal_I~~}		\sum_{i \in {E}}
	\end{array}		\end{array}
	</math>		</math>

	where <math>~~{\cal~~ E}</math> and <math>~~{\cal~~ I}</math> represent the sets of indices corresponding to the equality constraints and inequality constraints respectively, and <math>\lambda_i</math> is the Lagrange multiplier corresponding to constraint ''i''. In TOMLAB /KNITRO we define the feasibility error (Feas err) at a point ''xk ''to be the maximum violation of the constraints 6, 7, i.e., <math>\~~mbox~~{Feas err} = \max_{i \in ~~{\cal~~ E}\cup{~~\cal~~ I}~~} (~~0, \|h_i(x^k)\|, g_i(x^k)),</math> while the optimality error (Opt err) is defined as the maximum violation of the first two conditions 4, 5,		where <math>E</math> and <math>I</math> represent the sets of indices corresponding to the equality constraints and inequality constraints respectively, and <math>\lambda_i</math> is the Lagrange multiplier corresponding to constraint ''i''. In TOMLAB /KNITRO we define the feasibility error (Feas err) at a point ''xk ''to be the maximum violation of the constraints 6, 7, i.e., <math>\text{Feas err} = \max_{i} \in E \cup{I} 0, \|h_i(x^k)\|, g_i(x^k)),</math> while the optimality error (Opt err) is defined as the maximum violation of the first two conditions 4, 5,

	<math>		<math>
	\~~mbox~~{Opt err} = \max_{i \in ~~{\cal~~ I}}		\text{Opt err} = \max_{i} \in I
	(\\|\nabla_x {~~\cal~~ L}(x^k,\lambda^k) \\|_\infty, \|\lambda_i g_i(x^k)\|,		(\\|\nabla_x {L}(x^k,\lambda^k) \\|_\infty, \|\lambda_i g_i(x^k)\|,
	\|\lambda_i\|, \|g_i(x^k)\|).		\|\lambda_i\|, \|g_i(x^k)\|).
	</math>		</math>
Line 756:		Line 756:

	<math>		<math>
	\~~bea~~		\text{Feas err} \le \max (\tau_1 * \text{FEASTOL}, \text{FEASTOLABS})
	~~\mbox~~{Feas err} \le \max (\tau_1*\~~mbox{~~{~~\tt~~
	FEASTOL}}, \~~mbox~~{~~{\tt~~ FEASTOLABS}})
	</math>		</math>

	<math>		<math>
	\~~mbox~~{Opt err} \le \max (\tau_2*\mbox{~~{\tt~~ OPTTOL}},		\text{Opt err} \le \max (\tau_2*\mbox{OPTTOL}, \text{OPTTOLABS})
	\~~mbox~~{~~{\tt~~ OPTTOLABS}})
	~~\eea~~
	</math>		</math>

Line 948:		Line 944:

	<math>		<math>
	f(x) = \~~half~~ \sum_{j=1}^q r_j(x)^2.		f(x) = \frac{1}{2} \sum_{j=1}^q r_j(x)^2.
	</math>		</math>

Line 999:		Line 995:

	<math>		<math>
	\~~bea~~		\begin{array}{l}
	s_1 = c_1(x) \\		s_1 = c_1(x) \\
	s_2 = c_2(x) \\		s_2 = c_2(x) \\
	s_1 \times s_2 = 0, \quad s_1 \ge 0, \quad s_2 \ge 0.		s_1 \times s_2 = 0, \quad s_1 \ge 0, \quad s_2 \ge 0.
	\~~eea~~		\end{array}
	</math>		</math>

Elias: /* Overview */

2011-10-15T09:28:26Z

Overview

← Older revision		Revision as of 04:28, 15 October 2011
Line 1:		Line 1:
	==Overview==		==Introduction==

			===Overview===

	Welcome to the TOMLAB /KNITRO User's Guide. TOMLAB /KNITRO includes the KNITRO (MI)NLP solver from Ziena Optimization and an interface to The MathWorks' MATLAB.		Welcome to the TOMLAB /KNITRO User's Guide. TOMLAB /KNITRO includes the KNITRO (MI)NLP solver from Ziena Optimization and an interface to The MathWorks' MATLAB.

← Older revision		Revision as of 04:20, 20 December 2012
Line 273:		Line 273:
	\|''FEASTOL''\|\|Specifies the final relative stopping tolerance for the feasibility error. Smaller values of FEASTOL result in a higher degree of accuracy in the solution with respect to feasibility. Default 1''e - ''6.		\|''FEASTOL''\|\|Specifies the final relative stopping tolerance for the feasibility error. Smaller values of FEASTOL result in a higher degree of accuracy in the solution with respect to feasibility. Default 1''e - ''6.
	\|-valign="top"		\|-valign="top"
	\|''FEASTOL_ABS''\|\|Specifies the final absolute stopping tolerance for the feasibility error. Smaller values of FEASTOLABS result in a higher degree of accuracy in the solution with respect to feasibility. Default 0.		\|''FEASTOL_ABS''\|\|Specifies the final absolute stopping tolerance for the feasibility error. Smaller values of FEASTOLABS result in a higher degree of accuracy in the solution with respect to feasibility. Default 0. '''NOTE''': For more information on the stopping test used in TOMLAB /KNITRO see [http://tomwiki.com/KNITRO#Termination_Test_and_Optimality].
	\|-valign="tp"		\|-valign="tp"
	\|''GRADOPT''\|\|~~'''NOTE''': For more information on the stopping test used in TOMLAB /KNITRO see Section 5.~~		\|''GRADOPT''\|\|

	Specifies how to calculate first derivatives. In addition to the available numeric differentiation methods available in TOMLAB, KNITRO can internally estimate forward or centered numerical derivatives.		Specifies how to calculate first derivatives. In addition to the available numeric differentiation methods available in TOMLAB, KNITRO can internally estimate forward or centered numerical derivatives.
Line 289:		Line 289:
	3: KNITRO will estimate centered finite differential derivatives.		3: KNITRO will estimate centered finite differential derivatives.

	4: KNITRO expects exact first derivatives and performs gradient checking by ~~inter- nally~~ calculating forward finite differences.		4: KNITRO expects exact first derivatives and performs gradient checking by internally calculating forward finite differences.

	5: KNITRO expects exact first derivatives and performs gradient checking by ~~inter- nally~~ calculating centered finite differences.		5: KNITRO expects exact first derivatives and performs gradient checking by internally calculating centered finite differences.

	'''NOTE''': It is highly recommended to provide exact gradients if at all possible as this greatly impacts the performance of the code. For more information on these options see Section 8.1.		'''NOTE''': It is highly recommended to provide exact gradients if at all possible as this greatly impacts the performance of the code. For more information on these options see Section 8.1.
Line 305:		Line 305:
	4: KNITRO will compute Hessian-vector products using finite differences.		4: KNITRO will compute Hessian-vector products using finite differences.

	5: KNITRO expects the user to provide a routine to compute the Hessian-vector products. If this option is selected, the calculation is handled by the TOMLAB function ''~~nlp d2Lv~~.m''.		5: KNITRO expects the user to provide a routine to compute the Hessian-vector products. If this option is selected, the calculation is handled by the TOMLAB function ''nlp_d2Lv.m''.

	6: KNITRO will compute a limited-memory quasi-Newton BFGS.		6: KNITRO will compute a limited-memory quasi-Newton BFGS.
Line 341:		Line 341:
	\|''MAXTIMECPU''\|\|Specifies, in seconds, the maximum allowable CPU time before termination. Default 1''e''8.		\|''MAXTIMECPU''\|\|Specifies, in seconds, the maximum allowable CPU time before termination. Default 1''e''8.

	'''NOTE''': For more information on the stopping test used in TOMLAB /KNITRO see ~~Section 5~~.		'''NOTE''': For more information on the stopping test used in TOMLAB /KNITRO see [http://tomwiki.com/KNITRO#Termination_Test_and_Optimality].
	\|-valign="top"		\|-valign="top"
	\|''MAXTIMEREAL''\|\|Specifies, in seconds, the maximum allowable real time before termination. Default 1''e''8.		\|''MAXTIMEREAL''\|\|Specifies, in seconds, the maximum allowable real time before termination. Default 1''e''8.

	'''NOTE''': For more information on the stopping test used in TOMLAB /KNITRO see ~~Section 5~~.		'''NOTE''': For more information on the stopping test used in TOMLAB /KNITRO see [http://tomwiki.com/KNITRO#Termination_Test_and_Optimality].
	\|-valign="top"		\|-valign="top"
	\|''MIP_BRANCHRULE''\|\|Specifies which branching rule to use for MIP branch and bound procedure. De- fault 0.		\|''MIP_BRANCHRULE''\|\|Specifies which branching rule to use for MIP branch and bound procedure. De- fault 0.
Line 531:		Line 531:
	\|''OPTTOL_ABS''\|\|Specifies the final absolute stopping tolerance for the KKT (optimality) error. Smaller values of OPTTOLABS result in a higher degree of accuracy in the solution with respect to optimality. Default 0''.''0''e''0.		\|''OPTTOL_ABS''\|\|Specifies the final absolute stopping tolerance for the KKT (optimality) error. Smaller values of OPTTOLABS result in a higher degree of accuracy in the solution with respect to optimality. Default 0''.''0''e''0.

	'''NOTE''': For more information on the stopping test used in TOMLAB /KNITRO see ~~Section 5~~.		'''NOTE''': For more information on the stopping test used in TOMLAB /KNITRO see [http://tomwiki.com/KNITRO#Termination_Test_and_Optimality].
	\|-valign="top"		\|-valign="top"
	\|''OUTAPPEND''\|\|Specifies whether output should be started in a new file, or appended to existing files. The option affects knitro.log and files produced when "debug=1". It does not affect knitro newpoint.log, which is controlled by option "newpoint". Default: 0.		\|''OUTAPPEND''\|\|Specifies whether output should be started in a new file, or appended to existing files. The option affects knitro.log and files produced when "debug=1". It does not affect knitro newpoint.log, which is controlled by option "newpoint". Default: 0.