Documentation

ga

Find minimum of function using genetic algorithm

Syntax

x = ga(fitnessfcn,nvars)
x = ga(fitnessfcn,nvars,A,b)
x = ga(fitnessfcn,nvars,A,b,Aeq,beq)
x = ga(fitnessfcn,nvars,A,b,Aeq,beq,LB,UB)
x = ga(fitnessfcn,nvars,A,b,Aeq,beq,LB,UB,nonlcon)
x = ga(fitnessfcn,nvars,A,b,Aeq,beq,LB,UB,nonlcon,options)
x = ga(fitnessfcn,nvars,A,b,[],[],LB,UB,nonlcon,IntCon)
x = ga(fitnessfcn,nvars,A,b,[],[],LB,UB,nonlcon,IntCon,options)
x = ga(problem)
[x,fval] = ga(fitnessfcn,nvars,...)
[x,fval,exitflag] = ga(fitnessfcn,nvars,...)
[x,fval,exitflag,output] = ga(fitnessfcn,nvars,...)
[x,fval,exitflag,output,population] = ga(fitnessfcn,nvars,...)
[x,fval,exitflag,output,population,scores] = ga(fitnessfcn,nvars,...)

Description

x = ga(fitnessfcn,nvars) finds a local unconstrained minimum, x, to the objective function, fitnessfcn. nvars is the dimension (number of design variables) of fitnessfcn.

x = ga(fitnessfcn,nvars,A,b) finds a local minimum x to fitnessfcn, subject to the linear inequalities A*x ≤ b. ga evaluates the matrix product A*x as if x is transposed (A*x').

x = ga(fitnessfcn,nvars,A,b,Aeq,beq) finds a local minimum x to fitnessfcn, subject to the linear equalities Aeq*x = beq as well as A*x ≤ b. (Set A=[] and b=[] if no linear inequalities exist.) ga evaluates the matrix product Aeq*x as if x is transposed (Aeq*x').

x = ga(fitnessfcn,nvars,A,b,Aeq,beq,LB,UB) defines a set of lower and upper bounds on the design variables, x, so that a solution is found in the range LB  x  UB. (Set Aeq=[] and beq=[] if no linear equalities exist.)

x = ga(fitnessfcn,nvars,A,b,Aeq,beq,LB,UB,nonlcon) subjects the minimization to the constraints defined in nonlcon. The function nonlcon accepts x and returns vectors C and Ceq, representing the nonlinear inequalities and equalities respectively. ga minimizes the fitnessfcn such that C(x)  0 and Ceq(x) = 0. (Set LB=[] and UB=[] if no bounds exist.)

x = ga(fitnessfcn,nvars,A,b,Aeq,beq,LB,UB,nonlcon,options) minimizes with the default optimization parameters replaced by values in the structure options, which can be created using the gaoptimset function. (Set nonlcon=[] if no nonlinear constraints exist.)

x = ga(fitnessfcn,nvars,A,b,[],[],LB,UB,nonlcon,IntCon) requires that the variables listed in IntCon take integer values.

    Note:   When there are integer constraints, ga does not accept linear or nonlinear equality constraints, only inequality constraints.

x = ga(fitnessfcn,nvars,A,b,[],[],LB,UB,nonlcon,IntCon,options) minimizes with integer constraints and with the default optimization parameters replaced by values in the options structure.

x = ga(problem) finds the minimum for problem, where problem is a structure.

[x,fval] = ga(fitnessfcn,nvars,...) returns fval, the value of the fitness function at x.

[x,fval,exitflag] = ga(fitnessfcn,nvars,...) returns exitflag, an integer identifying the reason the algorithm terminated.

[x,fval,exitflag,output] = ga(fitnessfcn,nvars,...) returns output, a structure that contains output from each generation and other information about the performance of the algorithm.

[x,fval,exitflag,output,population] = ga(fitnessfcn,nvars,...) returns the matrix, population, whose rows are the final population.

[x,fval,exitflag,output,population,scores] = ga(fitnessfcn,nvars,...) returns scores the scores of the final population.

Input Arguments

fitnessfcn

Handle to the fitness function. The fitness function should accept a row vector of length nvars and return a scalar value.

When the 'Vectorized' option is 'on', fitnessfcn should accept a pop-by-nvars matrix, where pop is the current population size. In this case fitnessfcn should return a vector the same length as pop containing the fitness function values. fitnessfcn should not assume any particular size for pop, since ga can pass a single member of a population even in a vectorized calculation.

nvars

Positive integer representing the number of variables in the problem.

A

Matrix for linear inequality constraints of the form

A*x ≤ b.

If the problem has m linear inequality constraints and nvars variables, then

  • A is a matrix of size m-by-nvars.

  • b is a vector of length m.

ga evaluates the matrix product A*x as if x is transposed (A*x').

    Note:   ga does not enforce linear constraints to be satisfied when the PopulationType option is 'bitString' or 'custom'.

b

Vector for linear inequality constraints of the form

A*x ≤ b.

If the problem has m linear inequality constraints and nvars variables, then

  • A is a matrix of size m-by-nvars.

  • b is a vector of length m.

Aeq

Matrix for linear equality constraints of the form

Aeq*x = beq.

If the problem has m linear equality constraints and nvars variables, then

  • Aeq is a matrix of size m-by-nvars.

  • beq is a vector of length m.

ga evaluates the matrix product Aeq*x as if x is transposed (Aeq*x').

    Note:   ga does not enforce linear constraints to be satisfied when the PopulationType option is 'bitString' or 'custom'.

beq

Vector for linear equality constraints of the form

Aeq*x = beq.

If the problem has m linear equality constraints and nvars variables, then

  • Aeq is a matrix of size m-by-nvars.

  • beq is a vector of length m.

LB

Vector of lower bounds. ga enforces that iterations stay above LB. Set LB(i) = –Inf if x(i)is unbounded below.

    Note:   ga does not enforce bounds to be satisfied when the PopulationType option is 'bitString' or 'custom'.

UB

Vector of upper bounds. ga enforces that iterations stay below UB. Set UB(i) = Inf if x(i) is unbounded above.

nonlcon

Function handle that returns two outputs:

[c,ceq] = nonlcon(x)

ga attempts to achieve c ≤ 0 and ceq = 0. c and ceq are row vectors when there are multiple constraints. Set unused outputs to [].

You can write nonlcon as a function handle to a file, such as

nonlcon = @constraintfile

where constraintfile.m is a file on your MATLAB® path.

To learn how to use vectorized constraints, see Vectorized Constraints.

    Note:   ga does not enforce nonlinear constraints to be satisfied when the PopulationType option is set to 'bitString' or 'custom'.

    If IntCon is not empty, the second output of nonlcon (ceq) must be empty ([]).

    For information on how ga uses nonlcon, see Nonlinear Constraint Solver Algorithms.

options

Structure containing optimization options. Create options using gaoptimset, or by exporting options from the Optimization app as described in Importing and Exporting Your Work in the Optimization Toolbox™ documentation. Values in {} denote the default value. {}* means the default when there are linear constraints, and for MutationFcn also when there are bounds. You can also view the optimization parameters and defaults by typing gaoptimset at the command line. I* indicates that ga ignores or overwrites the option for mixed integer optimization problems.

OptionDescriptionValues

CreationFcn

I* Handle to the function that creates the initial population. See Population Options.

{@gacreationuniform} | {@gacreationlinearfeasible}* | Custom creation function

CrossoverFcn

I* Handle to the function that the algorithm uses to create crossover children. See Crossover Options.

{@crossoverscattered} for ga, {@crossoverintermediate}* for gamultiobj | @crossoverheuristic | @crossoversinglepoint | @crossovertwopoint | @crossoverarithmetic | Custom crossover function

CrossoverFraction

The fraction of the population at the next generation, not including elite children, that is created by the crossover function

Positive scalar | {0.8}

Display

Level of display

'off' | 'iter' | 'diagnose' | {'final'}

DistanceMeasureFcn

I* Handle to the function that computes distance measure of individuals, computed in decision variable or design space (genotype) or in function space (phenotype), for gamultiobj only

{@distancecrowding,'phenotype'}

EliteCount

Positive integer specifying how many individuals in the current generation are guaranteed to survive to the next generation. Not used in gamultiobj.

Positive integer | {ceil(0.05*PopulationSize)} | {0.05*(default PopulationSize)} for mixed-integer problems

FitnessLimit

Scalar. If the fitness function attains the value of FitnessLimit, the algorithm halts.

Scalar | {-Inf}

FitnessScalingFcn

Handle to the function that scales the values of the fitness function

Option unavailable for gamultiobj

{@fitscalingrank} | @fitscalingshiftlinear | @fitscalingprop | @fitscalingtop | Custom fitness scaling function

Generations

Positive integer specifying the maximum number of iterations before the algorithm halts

Positive integer |{100*numberOfVariables} for ga, {200*numberOfVariables} for gamultiobj

HybridFcn

I* Handle to a function that continues the optimization after ga terminates

or

Cell array specifying the hybrid function and its options structure

For gamultiobj, the only hybrid function is @fgoalattain

Function handle | @fminsearch | @patternsearch | @fminunc | @fmincon | {[]}

or

1-by-2 cell array | {@solver, hybridoptions}, where solver = fminsearch, patternsearch, fminunc, or fmincon {[]}

InitialPenalty

I* Initial value of penalty parameter

Positive scalar | {10}

InitialPopulation

Initial population used to seed the genetic algorithm; can be partial — see Population Options

Matrix | {[]}

InitialScores

I* Initial scores used to determine fitness; can be partial

Column vector | {[]}

MigrationDirection

Direction of migration — see Migration Options

'both' | {'forward'}

MigrationFraction

Scalar between 0 and 1 specifying the fraction of individuals in each subpopulation that migrates to a different subpopulation — see Migration Options

Scalar | {0.2}

MigrationInterval

Positive integer specifying the number of generations that take place between migrations of individuals between subpopulations — see Migration Options

Positive integer | {20}

MutationFcn

I* Handle to the function that produces mutation children. See Mutation Options.

{@mutationgaussian} for ga, {@mutationadaptfeasible}* for gamultiobj | @mutationuniform | Custom mutation function

NonlinConAlgorithm

Nonlinear constraint algorithm. See Nonlinear Constraint Solver Algorithms.

Option unchangeable for gamultiobj

{'auglag'} for ga, {'penalty'} for gamultiobj

OutputFcns

Functions that ga calls at each iteration. See Output Function Options.

Function handle or cell array of function handles | {[]}

ParetoFraction

I* Scalar between 0 and 1 specifying the fraction of individuals to keep on the first Pareto front while the solver selects individuals from higher fronts, for gamultiobj only

Scalar | {0.35}

PenaltyFactor

I* Penalty update parameter

Positive scalar | {100}

PlotFcns

Array of handles to functions that plot data computed by the algorithm. See Plot Options.

{[]} | @gaplotbestf | @gaplotbestindiv | @gaplotdistance | @gaplotexpectation | @gaplotgenealogy | @gaplotmaxconstr | @gaplotrange | @gaplotselection | @gaplotscorediversity | @gaplotscores | @gaplotstopping | Custom plot function

For gamultiobj there are additional choices: @gaplotpareto | @gaplotparetodistance | @gaplotrankhist | @gaplotspread

PlotInterval

Positive integer specifying the number of generations between consecutive calls to the plot functions

Positive integer | {1}

PopInitRange

Matrix or vector specifying the range of the individuals in the initial population. Applies to gacreationuniform creation function. ga shifts and scales the default initial range to match any finite bounds.

Matrix or vector | {[-10;10]} for unbounded components, {[-1e4+1;1e4+1]} for unbounded components of integer-constrained problems, {[lb;ub]} for bounded components, with the default range modified to match one-sided bounds.

PopulationSize

Size of the population

Positive integer | {50} when numberOfVariables <= 5, {200} otherwise | {min(max(10*nvars,40),100)} for mixed-integer problems

PopulationType

String describing the data type of the population — must be 'doubleVector' for mixed integer problems

'bitstring' | 'custom' | {'doubleVector'}

ga ignores all constraints when PopulationType is set to 'bitString' or 'custom'. See Population Options.

SelectionFcn

I* Handle to the function that selects parents of crossover and mutation children

Option unchangeable for gamultiobj

{@selectionstochunif} for ga, {@selectiontournament} for gamultiobj | @selectionremainder | @selectionuniform | @selectionroulette | Custom selection function

StallGenLimit

Positive integer. The algorithm stops if the average relative change in the best fitness function value over StallGenLimit generations is less than or equal to TolFun. If StallTest is 'geometricWeighted', then the algorithm stops if the weighted average relative change is less than or equal to TolFun.

Positive integer | {50} for ga, {100} for gamultiobj

StallTest

String describing the stopping test.

'geometricWeighted' | {'averageChange'}

StallTimeLimit

Positive scalar. The algorithm stops if there is no improvement in the objective function for StallTimeLimit seconds, as measured by tic and toc.

Option unavailable for gamultiobj

Positive scalar | {Inf}

TimeLimit

Positive scalar. The algorithm stops after running after TimeLimit seconds, as measured by tic and toc. This limit is enforced after each iteration, so ga can exceed the limit when an iteration takes substantial time.

Positive scalar | {Inf}

TolCon

Positive scalar. TolCon is used to determine the feasibility with respect to nonlinear constraints. Also, max(sqrt(eps),TolCon) determines feasibility with respect to linear constraints.

Positive scalar | {1e-3}

TolFun

Positive scalar. The algorithm stops if the average relative change in the best fitness function value over StallGenLimit generations is less than or equal to TolFun. If StallTest is 'geometricWeighted', then the algorithm stops if the weighted average relative change is less than or equal to TolFun.

Positive scalar | {1e-6} for ga, {1e-4} for gamultiobj

UseParallel

Compute fitness and nonlinear constraint functions in parallel, see Vectorize and Parallel Options (User Function Evaluation) and How to Use Parallel Processing.

true | {false}

Vectorized

Specifies whether functions are vectorized, see Vectorize and Parallel Options (User Function Evaluation) and Vectorize the Fitness Function.

'on' | {'off'}

IntCon

Vector of positive integers taking values from 1 to nvars. Each value in IntCon represents an x component that is integer-valued.

problem

Structure containing the following fields:

fitnessfcn

Fitness function

nvars

Number of design variables

Aineq

A matrix for linear inequality constraints

Bineq

b vector for linear inequality constraints

Aeq

Aeq matrix for linear equality constraints

Beq

beq vector for linear equality constraints

lb

Lower bound on x

ub

Upper bound on x

nonlcon

Nonlinear constraint function

rngstate

Optional field to reset the state of the random number generator

intcon

Index vector of integer variables

solver

'ga'

options

Options structure created using gaoptimset or the Optimization app

Create problem by exporting a problem from the Optimization app, as described in Importing and Exporting Your Work in the Optimization Toolbox documentation.

Output Arguments

x

Best point that ga located during its iterations.

fval

Fitness function evaluated at x.

exitflag

Integer giving the reason ga stopped iterating:

Exit FlagMeaning
1

Without nonlinear constraints — Average cumulative change in value of the fitness function over StallGenLimit generations is less than TolFun, and the constraint violation is less than TolCon.

With nonlinear constraints — Magnitude of the complementarity measure (see Complementarity Measure) is less than sqrt(TolCon), the subproblem is solved using a tolerance less than TolFun, and the constraint violation is less than TolCon.

2

Fitness limit reached and the constraint violation is less than TolCon.

3

Value of the fitness function did not change in StallGenLimit generations and the constraint violation is less than TolCon.

4

Magnitude of step smaller than machine precision and the constraint violation is less than TolCon.

5

Minimum fitness limit FitnessLimit reached and the constraint violation is less than TolCon.

0

Maximum number of generations Generations exceeded.

-1

Optimization terminated by an output function or plot function.

-2

No feasible point found.

-4

Stall time limit StallTimeLimit exceeded.

-5

Time limit TimeLimit exceeded.

When there are integer constraints, ga uses the penalty fitness value instead of the fitness value for stopping criteria.

output

Structure containing output from each generation and other information about algorithm performance. The output structure contains the following fields:

  • problemtype — String describing the type of problem, one of:

    • 'unconstrained'

    • 'boundconstraints'

    • 'linearconstraints'

    • 'nonlinearconstr'

    • 'integerconstraints'

  • rngstate — State of the MATLAB random number generator, just before the algorithm started. You can use the values in rngstate to reproduce the output of ga. See Reproduce Results.

  • generations — Number of generations computed.

  • funccount — Number of evaluations of the fitness function.

  • message — Reason the algorithm terminated.

  • maxconstraint — Maximum constraint violation, if any.

population

Matrix whose rows contain the members of the final population.

scores

Column vector of the fitness values (scores for integerconstraints problems) of the final population.

Examples

Given the following inequality constraints and lower bounds

[111221][x1x2][223],x10,  x20,

use this code to find the minimum of the lincontest6 function, which is provided in your software:

A = [1 1; -1 2; 2 1];
b = [2; 2; 3];
lb = zeros(2,1);
[x,fval,exitflag] = ga(@lincontest6,... 
    2,A,b,[],[],lb)

Optimization terminated: average change in
the fitness value less than options.TolFun.

x =
    0.6700    1.3310

fval =
   -8.2218

exitflag =
     1

Optimize a function where some variables must be integers:

fun = @(x) (x(1) - 0.2)^2 + ...
    (x(2) - 1.7)^2 + (x(3) - 5.1)^2;
x = ga(fun,3,[],[],[],[],[],[],[], ...
   [2 3]) % variables 2 and 3 are integers

Optimization terminated: average change
in the penalty fitness value less
than options.TolFun and constraint violation
is less than options.TolCon.

x =
    0.2000    2.0000    5.0000

Alternatives

For problems without integer constraints, consider using patternsearch instead of ga.

More About

expand all

Complementarity Measure

In the nonlinear constraint solver, the complementarity measure is the norm of the vector whose elements are ciλi, where ci is the nonlinear inequality constraint violation, and λi is the corresponding Lagrange multiplier.

Tips

  • To write a function with additional parameters to the independent variables that can be called by ga, see Passing Extra Parameters in the Optimization Toolbox documentation.

  • For problems that use the population type Double Vector (the default), ga does not accept functions whose inputs are of type complex. To solve problems involving complex data, write your functions so that they accept real vectors, by separating the real and imaginary parts.

Algorithms

For a description of the genetic algorithm, see How the Genetic Algorithm Works.

For a description of the mixed integer programming algorithm, see Integer ga Algorithm.

For a description of the nonlinear constraint algorithm, see Nonlinear Constraint Solver Algorithms.

References

[1] Goldberg, David E., Genetic Algorithms in Search, Optimization & Machine Learning, Addison-Wesley, 1989.

[2] A. R. Conn, N. I. M. Gould, and Ph. L. Toint. "A Globally Convergent Augmented Lagrangian Algorithm for Optimization with General Constraints and Simple Bounds", SIAM Journal on Numerical Analysis, Volume 28, Number 2, pages 545–572, 1991.

[3] A. R. Conn, N. I. M. Gould, and Ph. L. Toint. "A Globally Convergent Augmented Lagrangian Barrier Algorithm for Optimization with General Inequality Constraints and Simple Bounds", Mathematics of Computation, Volume 66, Number 217, pages 261–288, 1997.

Was this topic helpful?