Relaxation Heuristics - just solve an easier problem

Relaxation is a fundamental method for creating heuristics in difficult domains. This can allow us to tackle problems that were otherwise completely computationally intractable. Despite how important and impactful relaxation is, the idea itself is very intuitive and easy to understand. Join me as we begin to have a look at relaxation, learn that the common heuristics such as Manhattan Distance and Euclidean Distance were secretly relaxations the whole time, and learn how we can create our own. Watch the rest of the *Classical Planning Series*:    • Classical Planning (Masters of AI)   Need to understand the basics of search first? Check out my Search in a Nutshell course:    • Search in a Nutshell (bite-sized lessons c...   Watch the Pathfinding Fundamentals course:    • Pathfinding Fundamentals (Master's of AI: ...   Please consider subscribing!    / @lowerboundyt   This goes a long way to helping push me to continue making content like this. Thanks for watching! Chapters: 00:00 Intro 00:37 We need heuristics 01:28 Automatic & generalised 02:18 Design a simpler problem 03:18 Examples 03:43 Solve the relaxation 04:24 Computational tractability 06:03 How to relax a problem 08:41 Key properties 09:52 Outro