Since much of AI is "search", and search takes processing time, really the only thing you get from technology is being able to do either bigger, more elaborate searches, or do them faster. Technology only enables algorithms... it doesn't define them.

Put another way, the GOAP algorithm that Jeff wrote would simply be able to handle more possibilities in a shorter amount of time.

Fast pathfinding via Symmetry Breaking

The explicit identification and elimination of symmetries in pathfinding domains is an idea that until now has received little attention in the academic literature. Approaches for dealing with symmetry, such as Jump Point Search, provide us with powerful new tools for reducing the size of explicit regular search spaces. By eliminating symmetry we speed up not just A* but entire classes of similar pathfinding algorithms.

Also, consider: JPS is entirely orthogonal to almost every other speedup technique applicable to grid maps. Thus, there is no reason why we couldn't combine it, or other similar methods, with hierarchical pathfinding approaches, memory heuristics or even other optimality-preserving state-space reduction techniques. That means the results presented thus far are only the tip of the iceberg in terms of performant grid-based pathfinding methods.

From what Ive found out in my years of interest in AI - an awful lot of 'harder' AI is seaching through solution space (not just nice regular terrain style data, but irregularities of object analysis/manipulations and simulated behaviors).

Several difficulties :

With a complex environment to operate in, the number of interactions and possible interactions (and thus solutions) increases geometrically. Alot of data has to be sifted thru/evaluated to find that one thing you want to do. Along the same lines, the LOGIC (code) needed to handle the irrgularities of the environment grows geometrically also.

With a dynamic environment in a near realtime game (including the actions of other 'intelligent' objects) that changing environment must be RE-analyzed constantly (meaning you have to rerun much of the same analysis over and over frequently to be able to 'opportunisticly' interact.

One I found possibly of the highest difficulty is trying to come up with a unified measurement metric for the intelligent object to be able to decide between doing different possibilities (what is the priiority in that objects risk versus reward evaluations).

As usual having uncertainty (and being able to handle it) in the environment increases the difficulties at least a magnitude.

Fast pathfinding via Symmetry Breaking

Searched and found the word 'pre-processing' (as expected) on that which makes it a variant of portal analysis and hierarchical pathfinding.

Which is not a bad thing if it gave decent results when implemented in a game

Similar would be predigested influence maps to give an AI helpful data about good positions in a projectile type shooting/cover environment

Searched and found the word 'pre-processing' (as expected) on that which makes it a variant of portal analysis and hierarchical pathfinding.

The two Symmetry Breaking methods discussed, broadly classified as search space pruning techniques

Rectangular Symmetry Reduction

1. It preserves optimality.
2. It has a small memory overhead in practice.
3. Node insertion can be performed in constant time.
4. It can speed up A* search by anywhere from several factors to an order of magnitude.

Jump Point Search

1. It is optimal.
2. It involves no pre-processing.
3. It requires no extra-memory overheads.
4. It can consistently speed up A* search by over 10 times; making it not only competitive with, but often better than, approximate techniques such as HPA*

A* + JPS looks to be an old algorithm I considered long ago and only works effectively on simplistic maps -- no widely varying terrain costs (of passable terrain) and works best in largely open terrain (few small obstacles which can constantly short circuit the' jumps'). Thats fine within that limitation ("PS can speed up A* by a factor of between).

Something of interest would be where the crossover point is with 'typical' terrain (complexity) for the JPS method to wind up more costly than simple A* (or when possible hierarchical portal methods using A*)

The diagrams dont really illustrate how much backtracking can be needed when the current position has pocket culdesacs blocking the direct advancement towards the final goal. It looks like it starts doing alot of edge following when the jumps towards the target start failing, looking for a possible hole before backtracking away from the target.