I am working on a problem with data analytics. For the problem we are
dealing, we got stuck in dilemma whether to use A* or genetic algorithms.
Reason is, if we use A* algorithm we might get stuck in local optimum, but
if our heuristic function is using brute force approach (like, example:
comparing every possible sequence of data for what we are searching for), we
can reach to optimum solution but will take lots of computation and time. To
minimize time, we thought of using clusters of computers to crunch the data
and find the solution. This is a valid solution, but needs to spend more
computation power. Computation power can be optimized using few nodes in
clusters, when used genetic algorithms to solve the problem.
But here is the dilemma, are there any other benefits over other methods
using genetic algorithms, part from reducing computation time and power, and
of course finding optimum solution. Does genetic algorithms gives any edge
over other methods. Please let me know.
View this message in context: http://old.nabble.com/difference-between-A*-search-and-genetic-algorithms-tp34247203p34247203.html
Sent from the jgap-users mailing list archive at Nabble.com.
From: Jan Torben Heuer <mail@jt...> - 2012-09-18 11:23:33
On Friday 03 August 2012 09:41:57 sri-r-ksh wrote:
> if we use A* algorithm we might get stuck in local optimum
that should not happen if your heuristic function is correct. Do you get
correct results for h(x)=0 (Dijkstra)?
There is no such thing as "correct" heuristic function. There are different heuristic functions which behave well or not depending on the situation.
The need for heuristic functions is a difference between A* and genetic algorithms. Heuristic functions require extra domain knowledge. Genetic algorithms restrict domain knowledge to inside the fitness function and chromosomes.
> From: mail@...
> To: jgap-users@...
> Date: Tue, 18 Sep 2012 13:23:18 +0200
> Subject: Re: [jgap-users] difference between A* search and genetic algorithms
> On Friday 03 August 2012 09:41:57 sri-r-ksh wrote:
> > if we use A* algorithm we might get stuck in local optimum
> that should not happen if your heuristic function is correct. Do you get
> correct results for h(x)=0 (Dijkstra)?
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
> jgap-users mailing list