Results
Synthetic Problems
For synthetic benchmarks, we select two-objective scalable problems for comparison as described in Table 1. The ZDT test suite \cite{Zitzler_2000} provides a range of PF shapes, while the MW test suite \cite{Ma_2019} provides constraints and uniquely shaped PFs to challenge the optimisation algorithms. Both test suites rely on a similar construction method for minimization problems: taking a single variable function f1 against a shape function f2 as such:
\(\min\ f_1(x)=x_1\)
\(\min\ f_2(x)=g(x)h(f_1(x),g(x))\)
The single variable function closely resembles certain real-life multi-objective problems where an input is to be minimised against some other objective, for example minimizing process temperature, while achieving a target output. \cite{MacLeod_2022}