Paradox-free analysis for comparing the performance of optimization algorithms

Numerical comparison serves as a major tool in evaluating the performance of optimization algorithms, especially nondeterministic but existing methods may suffer from ‘cycle ranking’ paradox and/or ‘survival non-fittest’ paradox. This paper searches for paradox-free data analysis numerical comparison. It is discovered that class sufficient conditions exist designing analysis. Rigorous modeling and deduction are applied to profile employing filter. thus further proven algorithm-independent filter can prevent cycle ranking survival non-fittest paradoxes occurring. By adopting an filter, popular such ‘modified method’‘, accuracy method’, ‘the operational characteristics zones method’ be free comparing or benchmarking algorithms.