Overload basic arithmetic for classes arrays?

I have a an object obj with several fields, say

obj.a = [1 -1];
obj.b = [0 1; 0 0];

and can override a plus operator

obj = obj1 + obj2

as

obj.a = obj1.a + obj2.a
obj.b = obj1.b + obj2.b

The QUESTION 1: What if I have a class arrays, say obj1(1,2) and obj1(2,1) and want to sum them in the same way as it is performed for double i.e. getting obj(2,2)? Of course I need it for the arbitrary sizes and dimensions. Function arrayfun() is no help, since it does not do implicit expansion. The bsxfun() would do the job, but it is defined for double, int, char, etc.

The QUESTION 2: Same question for mtimes: how can i do matrix multiplication if times function is defined?

Answer 1

[Deepak]

Hi @Alex Kashuba,

The task of overloading the “plus” operator for a custom MATLAB class can be achieved by defining a method named “plus” within the class that dictates how two instances of the class should be added.

First, determine the sizes of the input object arrays and compute the resulting size using “max(size(obj1), size(obj2))” function. Next, expand the input arrays to this common size using “repmat” function, ensuring that each element aligns correctly for addition. Finally, iterate over the expanded arrays, performing the addition operation on each corresponding element's properties, and return a new array of the custom objects containing the summed results.

Here is the MATLAB code to achieve the same:

classdef MyObject 
    properties 
        a 
        b 
    end 
    
    methods 
        function obj = MyObject(a, b) 
            if nargin > 0 
                obj.a = a; 
                obj.b = b; 
            end 
        end 
        
        function obj = plus(obj1, obj2) 
            
            size1 = size(obj1); 
            size2 = size(obj2); 
            
            % Determine the resulting size using implicit expansion rules 
            resultSize = max(size1, size2); 
            
            % Expand obj1 and obj2 to the resulting size 
            obj1 = repmat(obj1, resultSize ./ size1); 
            obj2 = repmat(obj2, resultSize ./ size2); 
            
            % Perform element-wise addition 
            obj(resultSize(1), resultSize(2)) = MyObject(); 
            for i = 1:resultSize(1) 
                for j = 1:resultSize(2) 
                    obj(i, j).a = obj1(i, j).a + obj2(i, j).a; 
                    obj(i, j).b = obj1(i, j).b + obj2(i, j).b; 
                end 
            end 
        end 
    end 
end 
%Test Script 
obj1(1, 1) = MyObject([1 -1], [0 1; 0 0]); 
obj1(1, 2) = MyObject([2 2], [1 0; 0 1]); 

obj2(1, 1) = MyObject([3 3], [1 1; 1 1]); 
obj2(2, 1) = MyObject([0 -1], [0 0; 1 0]); 

% Perform addition with implicit expansion 
sumObj = obj1 + obj2; 

For more information on the functions used refer to the following documentation:

Classes: www.mathworks.com/help/matlab/object-oriented-programming.html

repmat: www.mathworks.com/help/matlab/ref/repmat.html

I hope this assists in resolving the issue.