I just swapped in the Jenetics 4.1.0 libraries for version 4.0.0. I'm now getting a NULL result here occasionally (random seed dependent) when I use Limits.byPopulationConvergence():

        final EvolutionResult<DoubleGene, Double> result = engine.stream()
                .limit(Limits.byPopulationConvergence(.03))
                .limit(maxGeneration)
                .peek(this::logger)
                .collect(toBestEvolutionResult());

If I reduce the argument to Limits.byPopulationConvergence() enough, it runs every time. This is a toy problem that may be converged at the first generation. Similar problems with more diversity don't have this problem.