Matlab-使用arrayfun评估嵌套匿名函数

Question

I want to evaluate an anonymous function in Matlab, which again consists of several anonymous functions.

Please excuse the somewhat chaotic labeling of the variables as this is a strongly simplified version of the original problem:

C.B_MT = 2; C.T = 24; C.OM_MT = 1/24; C.P_WTmax = 2;
ub = 3;
ObjFcn = @(x) CostFcn1(x,C,ub);
x = 1:4;
fit = ObjFcn(x)

The function that ObjFcn calls is the following:

function F = CostFcn1(x,C,ub)
    F = f_t(x,C,ub) + OM_DG(C) + TCPD_BES(x,C);
    function ft = f_t(x,C,ub)
        N = cell(1,2);
        ft = x(2)*C.B_MT+ub;
        for i = 1:2
            N{i} = C.B_MT*x(i+2)+ub;
        end
        ft = ft(x) + sum(arrayfun(@(N_q) N_q{1}(x), N));
    end

    function om_dg = OM_DG(C)
        om_dg = C.T*C.OM_MT;
    end

    function tcpd = TCPD_BES(x,C)
        tcpd = x(1)*C.P_WTmax;
    end
end

When I run the code without the part of the for-loop and arrayfun , it works without a problems ( fit = 10 ). With arrayfun , however, I get an error. How can I change the code that arrayfun doesn't cause any problems?

Answer 1

The problem is with the part where you are trying to access ft(x) and N_q{1}(x) . Replace these parts with ft and N_q{1} and you should be good to go.

Also, unless you need to do anything else in the three nested functions you have. you can get rid of these functions and just directly use the statements inside of them. that will save you a lot of overhead. Here I add the optimized code:

function F = CostFcn1(x,C,ub)
    om_dg = C.T*C.OM_MT;
    tcpd = x(1)*C.P_WTmax;
    N = zeros(1,2);
    ft = x(2)*C.B_MT+ub;
    for i = 1:2
        N(i) = C.B_MT*x(i+2)+ub;
    end
    ft = ft + sum(N);
    F = ft + om_dg + tcpd;
end

This should work if your original code is similar (the variables used here have the same type as in your original code) to the one you have posted