What Are the Heuristic Search Techniques?
Heuristics is a problem-solving method that trades comprehensiveness for efficiency. It uses certain methodologies to zero in quickly on an adequate (though not necessarily optimal) solution. Heuristics has broad applications in mathematics, science and other fields involving abstract problem-solving.
Other People Are Reading
-
Hill Climbing
-
"Hill climbing" is a classic heuristic technique that uses the metaphor of climbing a hill. The first step involves determining if the initial state is the goal state itself; if not, it becomes the current state. The next step involves advancing one state and evaluating if it is approaching the goal state, and continuing in this fashion until reaching a local optimum. Under this method, a plateau means that all neighboring states (states which are the same degree of separation from the goal) are identical.
Algorithmic Heuristic Searching
-
Algorithms can also be applied to heuristic search techniques. For example, in the function f(x) = g(x) + h'(x), x represents the goal state, h'(x) is the expected cost involved in advanced from the current state to goal and g(x) is the cost required to advance from the initial state to the heuristic x. Algorithmic problem solving serves to minimize the path cost while searching for a solution in the least amount of time.
-
Simulated Annealing
-
Similar to the hill method, the technique known as "simulated annealing" is itself a metaphor based on the metallurgical process of heating and then cooling a metal to achieve the ideal state of crystallization. Applied to heuristics, this means permitting large "downhill" jumps that may be "worse" (farther from the goal state) in order to avoid plateaus or other heuristic dead ends.
Steep Ascent Hill Climbing
-
This is a variation on the conventional hill climbing technique. It's also referred to as gradient searching. The steep ascent technique requires one to advance to the optimal state that is a single move away. When one move is chosen, the remaining available moves are rejected and not revisited, so the steep ascent resembles a continuous, unbroken line along the "steepest slope."
-
