Heuristic Search之Greedy Best First Search

本篇文章介绍Greedy Best Fisrt Search算法。实现GBFS算法时使用了上篇文章介绍的优先队列。下篇文章将尝试介绍著名的A star算法。

Greedy Best-First Search

• use heuristic function as evaluation function: f(n) = h(n)
• always expands the node that is closest to the goal node
• eats the largest chunk out of the remaining distance, hence, “greedy”

The following example is “Touring in Romania”, which is an actual problem for making a plan travelling from Arad to Bucharest, the aim that we use the lowest cost.

example

Touring in Romania: Heuristic

• HSLD(n) = straight-line distance to Bucharest straight-line distance: Euclidean distance
• distance to Bucharest because our goal is to be in Bucharest

[table]

• HSLD(Bucharest) = 0
• HSLD(Fagaras) = 176 < 211 driving distance
• HSLD(n) cannot be computed from the problem description, it represents
  additional information

What’s wrong with GBFS?

for this problem: search proceeds straight to the goal node:

• minimal search cost
• but not the optimal path

Compare to Uniform-Cost Search

• UCS: numbers start from 0 and increase – tendency to expand earlier nodes –
  breadth-first tendency
• UCS的介绍
• GBFS: number start from high and decreases – tendency to expand later nodes
  – depth-first tendency

参考资料

1. 数据结构与算法分析
2. 大话数据结构
3. AI planning 课程资料
4. AI planning Vedio