【ALGO】遗传算法(2)

未成熟收敛问题

未成熟收敛现象

当遗传算法没有找到最优解或者满意解时，群体中不能个体到达非常相似的情况，特征是群体中个体结构的多样性急剧减少.

原因

• 选择，交叉，变异等操作并不能搜索到全部解空间
• 存在随机误差(取样误差和选择误差)
• 求解问题可能是GA欺骗问题

解决方案

• 重新启动法
• 匹配策略(Mating Strategies)
• 重组策略(Recombination Strategies)
• 替代策略(Replacement Strategies)

GA终止规则

• 给定一个最大遗传代数MaxGen，算法迭代到MaxGen时算法停止.
• 当进化两代中最优个体小于要求的偏差时，算法终止
• 所有个体或者指定比例以上个体趋同时，要求停止计算
• 算法迭代达到最大时间限制

Matlab中GA接口

x = ga(fun, nvars, A, b); 

finds a local minimum x to fun, subject to the linear inequalities A*x<=b. ga evaluates the matrix A*x as if x is transposed A*x'.

x = ga(fun, nvars, A, b, Aeq, beq);

finds a local minimum x to fuin, subject to the linear equalities Aeq*x=beq and 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(fun, 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.

x = ga(fun, 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 fun such as C(x)<=0 and Ceq(x)=0.

x=ga(fun, nvars, A, b, Aeq, lb, ub, nonlcon, options);

minimizes with the default optimization parameters replaced by values in options. Create options using optimoptions.

x=ga(fun, 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 inequalitiy constraints.

xi=linspace(-6, 2, 300);
yi=linspace(-4, 4, 300);
[X, Y]=meshgrid(xi, yi);
Z=ps_example([X(:), Y(:)]);
Z=reshape(Z, size(X));
surf(X, Y, Z, 'MeshStyle', 'none');
colormap 'jet';
view(-26, 43);
xlabel('x(1)');
ylabel('x(2)');
title('ps\_example(x)');

Use the genetic algorithm to minimize the ps_example function

rng default
x = ga(@ps_example, 2);

Minimize a Nonsmooth function with linear constraints，set the region is x(1)+x(2)>=1 and x(2)<=5+x(1). Convert the two inequality constraints to the matrix form A*x<=b. In other words, get the x variables on the left-hand side of the inequality, and make both inequalities less than or equal.

A = [-1, -1; -1, 1];
b = [-1; 5];
x_2 = ga(@ps_example, 2, A, b);

Minimize a Nonsmooth function with linear equality and inequality constraints，set the inequalities and equalities as follows

x(1)+x(2)>=1
x(2)==5+x(1)

rng default
A = [-1 -1];
b = -1;
Aeq = [-1 1];
beq = 5;
x_3 = ga(@ps_example, 2, A, b, Aeq, beq);

Optimize with Linear Constraints and Bounds，set bounds 1<=x(1)<=6 and -3<=x(2)<=8.

A = [-1 -1];
b = -1;
Aeq = [-1 1];
beq = 5;
lb = [1 -3];
ub = [6 8];
x_4 = ga(@ps_example, 2, A, b, Aeq, beq, lb, ub);

Optimize with Nonlinear using ga, set the region as
$$\{\begin{array}{l}2x_1^2+x_2^2\le 3\\ (x_1+1{)}^2=(x_2/2{)}^4\end{array}$$
function ellipsecons is as follows

function [c, ceq] = ellipsecons(x)
    c = 2*x(1)^2+x(2)^2-3;
    ceq = (x(1)+1)^2-(x(2)/2)^4;
end

和

x_5 = ga(@ps_example, 2, [], [], [], [], [], [], @ellipsecons);

Minimize with nondefault options, use the ga to minimize function on the region
$$\{\begin{array}{l}-x_1-x_2\le -1\\ -x_1+x_2==5\end{array}$$
To obtain a more accurate solution, set a constraint tolerance of 1e-6. And to monitor the solver progress, set a plot function.

options = optimoptions('ga', 'ConstraintTolerance', 1e-6, 'PlotFcn', @gaplotbestf);
x_6 = ga(@ps_example, 2, A, b, Aeq, beq, [], [], [], options);

Minimize a nonlinear function with Integer Constraints, use ga to minimize an intger-constrained nonlinear problem.

IntCon = 1;
fun = @ps_example;
A = [];
b = [];
Aeq = [];
beq = [];
lb = [];
ub = [];
nonlcon = [];
[x_7, fval] = ga(fun, 2, A, b, Aeq, beq, lb, ub, nonlcon, IntCon);

Obtain Diagnostic information, constrained to have $x_1$ integer-valued.

IntCon = 1;
fun = @ps_example;
A = [];
b = [];
Aeq = [];
beq = [];
lb = [];
ub = [];
nonlcon = [];
options = optimoptions('ga', 'PlotFcn', @gaplotbestf);
[x_8, fval, exitflag, output] = ga(fun, 2, A, b, Aeq, beq, lb, ub, nonlcon, IntCon, options);

Reason that ga stopped, returned as an integer

参考资料

Matlb Global Optimization Toolbox