Algorithm 2 Multi-objective evolutionary strategy.
for step ← to max_steps do
token, solution ← sampler.ask()	▹ ask for candidate solution
loss ← evaluate(solution)	▹ evaluate the given solution
sampler.update(token, loss)	▹ give fitness result back to ES
end for
results ← sampler.results()
first_front ← get_first_pareto_front(results)