Hi Peter,
"fmincon()" is a function in the MATLAB Optimization Toolbox used to optimize constrained nonlinear multivariable function.
As the document states, "scale in some sense with problem size or values", it means that the tolerance values can be adjusted automatically based on the characteristics of the optimization problem. The scaling depends on the type of algorithm used by "fmincon()" as well as magnitude of the values of the objective function and the constraints. More information on this can be found in "Tolerances by Solver and Algorithm" in the documentation: https://www.mathworks.com/help/optim/ug/tolerance-details.html
For eg, if the problem involves optimizing a function with very large values, the tolerance constraints might be scaled to a larger absolute value to reflect the magnitude of the function values. Similarly, if the problem has a large number of variables or constraints, the tolerance values might be scaled down to account for the increased complexity of the problem.
