Heuristics vs Optimal Algorithms

The cluster debates the use of heuristics, approximations like genetic algorithms and simulated annealing, versus exact or optimal algorithms for solving hard optimization problems, questioning the feasibility and benefits of each approach.

➡️ Stable 0.7x AI & Machine Learning
3,974
Comments
20
Years Active
5
Top Authors
#2765
Topic ID

Activity Over Time

2007
14
2008
34
2009
61
2010
108
2011
142
2012
132
2013
110
2014
149
2015
184
2016
202
2017
279
2018
242
2019
244
2020
258
2021
263
2022
340
2023
315
2024
462
2025
400
2026
35

Keywords

GMRES TSP MCMC ML SGD NOT UCT YES ycombinator.com heuristics search algorithm heuristic optimal greedy approach exhaustive instruction solution

Sample Comments

WithinReason Mar 1, 2023 View on HN

How many algorithms don't benefit from optimisations you think?

thepapanoob Nov 13, 2018 View on HN

another huge thing is it will be really hard to get optimized algorithms

0xFACEFEED Jul 10, 2022 View on HN

A heuristics based approach like yours is very underrated.

ggm Jul 6, 2023 View on HN

How about a heuristic instead of an algorithm.

DasIch Jun 13, 2016 View on HN

Why rely on a heuristic, when you can do better?

fiddlerwoaroof Sep 22, 2023 View on HN

It’s still an optimization problem, and the results might be surprising

gameswithgo Jan 5, 2020 View on HN

sounds like a decent simulated annealing or genetic algorithm search. which seems fine. smart people trying different ideas for a while.

LaserToy Jan 19, 2019 View on HN

Bfs is just one of examples. I can ask you a question you will not be able to answer without knowing the answer. There is a reason it took researchers years to find those optimal solutions

raverbashing Sep 29, 2014 View on HN

Yes, yes it isHeuristics may help, but in the end you either go through "all" of it or use approximate methods that will give you a good solution but not necessarily the best one (like Genetic Algorithms, Simulated Annealing, swarm optimization, etc)

agentultra Apr 11, 2025 View on HN

Some heuristics are better than others.