MATLAB OPTIMIZATION TOOLBOX - RELEASE NOTES Manuel d'utilisateur

Optimization Toolbox™ Release Notes

x ContentsR2006a Bug FixesR14SP3 Notify Parameter Added to Display Option for FiveFunctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

R2015aVersion: 7.2New FeaturesBug FixesCompatibility Considerations

R2015a1-2Improved performance and robustness of intlinprog primal-simplex algorithmThe 'primal-simplex' algorithm of the intlinprog solver c

1-3Compatibility ConsiderationsAny code that checks for this error will have different behavior than before.Least-squares solvers allow equal upper a

R2014bVersion: 7.1New FeaturesBug FixesCompatibility Considerations

R2014b2-2Dual-simplex algorithm in linprog linear programming solverlinprog has a new algorithm option named 'dual-simplex'. The algorithm s

2-3That method still works. But when using optimoptions, you cannot set the parameterin the Algorithm option, and instead must use the InitDamping op

R2014b2-4bintprog removedThe bintprog function has been removed.Compatibility ConsiderationsTo update code to use intlinprog instead of bintprog, see

R2014aVersion: 7.0New FeaturesBug FixesCompatibility Considerations

How to Contact MathWorksLatest news:www.mathworks.comSales and services:www.mathworks.com/sales_and_servicesUser community:www.mathworks.com/matlabcen

R2014a3-2Mixed-integer linear programming solverThe intlinprog function solves mixed-integer linear programming problems. For moreinformation, see the

3-3Nonlinear least-squares tweaksThe lsqcurvefit and lsqnonlin solvers have slightly different behavior than beforewhen the internally calculated tru

R2013bVersion: 6.4New FeaturesBug FixesCompatibility Considerations

R2013b4-2Solvers that check initial point more carefullyAll nonlinear solvers now check whether derivatives of the objective and nonlinearconstraint f

R2013aVersion: 6.3New FeaturesBug FixesCompatibility Considerations

R2013a5-2optimoptions function for setting options with compact andcomprehensive displayThe new optimoptions function creates and modifies options for

5-3Solver Previous Setting Current Setting'LargeScale' = 'off' 'Algorithm' = 'quasi-newton''LargeScale&a

R2012bVersion: 6.2.1Bug FixesCompatibility Considerations

iiiContentsR2015a Improved performance and robustness of intlinprogprimal-simplex algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2lin

R2012b6-2Changing default algorithms for fmincon and quadprogThe fmincon and quadprog default algorithms will change in a future release.• The default

R2012aVersion: 6.2New FeaturesBug FixesCompatibility Considerations

R2012a7-2Enhanced Robustness in fminuncThe fminunc medium-scale algorithm now attempts to recover from failures whenevaluating the objective function

7-3Compatibility ConsiderationsThe fmincon sqp algorithm can give slightly different results than before.

R2011bVersion: 6.1New FeaturesBug FixesCompatibility Considerations

R2011b8-2Derivative Estimate Changes• The fsolve, lsqcurvefit, and lsqnonlin solvers now accept the FinDiffType option. SetFinDiffType to 'centra

8-3Solvers no longer use the LevenbergMarquardt, LineSearchType, andNonlEqnAlgorithm options, since these options relate only to the Gauss-Newtonalgo

R2011b8-4is 'on'). If you provide gradients, have no linear constraints, and want to obtain theprevious behavior, set ScaleProblem to '

8-5exception = MException.last;MSGID = exception.identifier;Tip Warning messages indicate a potential issue with your code. While you can turn off aw

iv ContentsPlot functions and output functions for monitoring progressof intlinprog solver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

R2011aVersion: 6.0New FeaturesBug FixesCompatibility Considerations

R2011a9-2New Quadratic Programming Algorithmquadprog has a new algorithm named 'interior-point-convex'. It has thesefeatures:• The algorithm

9-3Enhanced Robustness in Nonlinear SolversMore solvers now attempt to recover from errors in the evaluation of objective functionsand nonlinear cons

R2011a9-4Output Structure TweakFor the trust-region-reflective algorithm, the algorithm field of the outputstructure is now 'trust-region-reflect

R2010bVersion: 5.1New FeaturesBug Fixes

R2010b10-2Enhanced fmincon Finite Difference Algorithms Add RobustnessThe fmincon interior-point and sqp algorithms now attempt to recover from errors

R2010aVersion: 5.0New FeaturesBug FixesCompatibility Considerations

R2010a11-2New fmincon Algorithmfmincon has a new algorithm called SQP for Sequential Quadratic Programming. Thealgorithm has the following features:•

R2009bVersion: 4.3New FeaturesBug FixesCompatibility Considerations

vR2013b Solvers that check initial point more carefully . . . . . . . . . . . 4-2R2013a optimoptions function for setting options with compact andcomp

R2009b12-2Enhanced Exit Messages in Selected SolversEnhanced, clearer exit messages in fsolve, lsqnonlin, and lsqcurvefit, with linksfor more informat

12-3The linprog output structure now contains the constrviolation field, which reportsthe maximum constraint function at the final point.Compatibilit

R2009aVersion: 4.2New FeaturesBug FixesCompatibility Considerations

R2009a13-2Parallel Gradient Estimation Available in fmincon Interior-Point AlgorithmThe fmincon solver's interior-point algorithm can now compute

13-3• 'iter-detailed'• 'notify-detailed'These settings have the same effect as the corresponding settings without '-detailed

R2009a13-4Compatibility ConsiderationsThis change might cause users (or code) that examine exit flags to evaluate a resultmore favorably than previous

R2008bVersion: 4.1New FeaturesBug FixesCompatibility Considerations

R2008b14-2fsolve, lsqcurvefit, lsqnonlin Algorithm and Options Changes• The Levenberg-Marquardt algorithm was refactored in the solvers fsolve, lsqcur

14-3• fmincon• fminimax• fminunc• fseminfThe fmincon active-set algorithm and fminunc medium-scale algorithm gained centralfinite differences this re

vi ContentsR2011b Derivative Estimate Changes . . . . . . . . . . . . . . . . . . . . . . . . . . 8-2Gauss-Newton Algorithm Removed . . . . . . . . .

R2008b14-40. Forward finite differences are typically calculated with a step size proportional tosqrt(eps), which is about 1.5*10–8. When a component

R2008aVersion: 4.0New FeaturesBug FixesCompatibility Considerations

R2008a15-2Parallel Computing Toolbox Support in fmincon, fminimax, andfgoalattainfmincon, fminimax, and fgoalattain can take finite differences in par

15-3= 'off' and Algorithm = 'trust-region-reflective'. In this case, the 'active-set' algorithm is used, to minimize ba

R2007bVersion: 3.1.2Bug Fixes

R2007aVersion: 3.1.1New FeaturesBug FixesCompatibility Considerations

R2007a17-2Changes to Outputs of Multiobjective Solvers• fminimax now returns the value of max(fval) in the output maxfval.• The iterative display of f

R2006bVersion: 3.1New FeaturesBug FixesCompatibility Considerations

viiktrlink Available for Macintosh 64-Bit Systems . . . . . . . . . . 10-2Output Structure Tweaks . . . . . . . . . . . . . . . . . . . . . . . . . .

R2006b18-2New Optimization ToolThe Optimization Tool is a graphical user interface (GUI) for performing commonoptimization tasks with the Optimization

18-3Compatibility ConsiderationsThe above changes to the input x and optimValues structure have the followingcompatibility considerations in the outp

R2006aVersion: 3.0.4Bug Fixes

R14SP3Version: 3.0.3New FeaturesBug Fixes

R14SP320-2Notify Parameter Added to Display Option for Five FunctionsYou can now set the optimization option Display to 'notify' for the fun

viii ContentsEnhanced Exit Messages in Selected Solvers . . . . . . . . . . . . 13-2Links to More Information Window . . . . . . . . . . . . . . . . .

ixNew fmincon Solver, New Option Algorithm for fmincon,Option LargeScale Changed . . . . . . . . . . . . . . . . . . . . . . . . 15-2External Interfac