Aren't these all fitness scoring methods for what GA calls mutation, crossover, and selection?
And do the functions under test stably coverage on the same output even if not exactly functionally isomorphic? When are less costly approximate solutions sufficient?
Big O, Dynamic tracing and instrumentation, costed opcodes, OPS/kWHr/$
Re: costed opcodes in BPF, eWASM (like EVM) https://news.ycombinator.com/item?id=40475509 :
Aren't these all fitness scoring methods for what GA calls mutation, crossover, and selection?
And do the functions under test stably coverage on the same output even if not exactly functionally isomorphic? When are less costly approximate solutions sufficient?
Big O, Dynamic tracing and instrumentation, costed opcodes, OPS/kWHr/$
https://news.ycombinator.com/item?id=41333249 :