Documentation |
Suppose you want to minimize the simple fitness function of two variables x_{1} and x_{2},
$$\underset{x}{\mathrm{min}}f(x)=100{\left({x}_{1}^{2}-{x}_{2}\right)}^{2}+{(1-{x}_{1})}^{2}$$
subject to the following nonlinear inequality constraints and bounds
$$\begin{array}{ll}{x}_{1}\cdot {x}_{2}+{x}_{1}-{x}_{2}+1.5\le 0\hfill & \text{(nonlinearconstraint)}\hfill \\ 10-{x}_{1}\cdot {x}_{2}\le 0\hfill & \text{(nonlinearconstraint)}\hfill \\ 0\le {x}_{1}\le 1\hfill & \text{(bound)}\hfill \\ 0\le {x}_{2}\le 13\hfill & \text{(bound)}\hfill \end{array}$$
Begin by creating the fitness and constraint functions. First, create a file named simple_fitness.m as follows:
function y = simple_fitness(x) y = 100*(x(1)^2 - x(2))^2 + (1 - x(1))^2;
(simple_fitness.m ships with Global Optimization Toolbox software.)
The genetic algorithm function, ga, assumes the fitness function will take one input x, where x has as many elements as the number of variables in the problem. The fitness function computes the value of the function and returns that scalar value in its one return argument, y.
Then create a file, simple_constraint.m, containing the constraints
function [c, ceq] = simple_constraint(x) c = [1.5 + x(1)*x(2) + x(1) - x(2);... -x(1)*x(2) + 10]; ceq = [];
(simple_constraint.m ships with Global Optimization Toolbox software.)
The ga function assumes the constraint function will take one input x, where x has as many elements as the number of variables in the problem. The constraint function computes the values of all the inequality and equality constraints and returns two vectors, c and ceq, respectively.
To minimize the fitness function, you need to pass a function handle to the fitness function as the first argument to the ga function, as well as specifying the number of variables as the second argument. Lower and upper bounds are provided as LB and UB respectively. In addition, you also need to pass a function handle to the nonlinear constraint function.
ObjectiveFunction = @simple_fitness; nvars = 2; % Number of variables LB = [0 0]; % Lower bound UB = [1 13]; % Upper bound ConstraintFunction = @simple_constraint; rng(1,'twister') % for reproducibility [x,fval] = ga(ObjectiveFunction,nvars,... [],[],[],[],LB,UB,ConstraintFunction) Optimization terminated: average change in the fitness value less than options.TolFun and constraint violation is less than options.TolCon. x = 0.8123 12.3104 fval = 1.3574e+04
The genetic algorithm solver handles linear constraints and bounds differently from nonlinear constraints. All the linear constraints and bounds are satisfied throughout the optimization. However, ga may not satisfy all the nonlinear constraints at every generation. If ga converges to a solution, the nonlinear constraints will be satisfied at that solution.
ga uses the mutation and crossover functions to produce new individuals at every generation. ga satisfies linear and bound constraints by using mutation and crossover functions that only generate feasible points. For example, in the previous call to ga, the mutation function mutationguassian does not necessarily obey the bound constraints. So when there are bound or linear constraints, the default ga mutation function is mutationadaptfeasible. If you provide a custom mutation function, this custom function must only generate points that are feasible with respect to the linear and bound constraints. All the included crossover functions generate points that satisfy the linear constraints and bounds except the crossoverheuristic function.
To see the progress of the optimization, use the gaoptimset function to create an options structure that selects two plot functions. The first plot function is gaplotbestf, which plots the best and mean score of the population at every generation. The second plot function is gaplotmaxconstr, which plots the maximum constraint violation of nonlinear constraints at every generation. You can also visualize the progress of the algorithm by displaying information to the command window using the 'Display' option.
options = gaoptimset('PlotFcns',{@gaplotbestf,@gaplotmaxconstr},'Display','iter');
Rerun the ga solver.
[x,fval] = ga(ObjectiveFunction,nvars,[],[],[],[],... LB,UB,ConstraintFunction,options) Best max Stall Generation f-count f(x) constraint Generations 1 2670 13666.1 0 0 2 5282 13578.2 6.648e-05 0 3 7894 14043.7 0 0 4 13156 14045 0 0 5 17806 13573.5 0.000999 0 Optimization terminated: average change in the fitness value less than options.TolFun and constraint violation is less than options.TolCon. x = 0.8122 12.3103 fval = 1.3573e+04
You can provide a start point for the minimization to the ga function by specifying the InitialPopulation option. The ga function will use the first individual from InitialPopulation as a start point for a constrained minimization.
X0 = [0.5 0.5]; % Start point (row vector) options = gaoptimset(options,'InitialPopulation',X0);
Now, rerun the ga solver.
[x,fval] = ga(ObjectiveFunction,nvars,[],[],[],[],... LB,UB,ConstraintFunction,options) Best max Stall Generation f-count f(x) constraint Generations 1 2670 13578.1 0.0005253 0 2 5282 13578.2 1.479e-05 0 3 8394 14037.2 0 0 4 17094 13573.6 0.0009898 0 Optimization terminated: average change in the fitness value less than options.TolFun and constraint violation is less than options.TolCon. x = 0.8123 12.3103 fval = 1.3574e+04