Working backward from the goal

From Wikipedia, the free encyclopedia

Working backward from the goal is an application of the backward chaining strategy in artificial intelligence. When problem solving, working backward is one way to reach the goal of solving a problem. George Pólya's book How to Solve It lists working backward in the section on problem solving strategies.

Contents 1 Primates do it 2 Applications to theorem-proving 3 Applications to AI 4 See also 5 Notes

[edit] Primates do it

Kanzi, a bonobo (pygmy chimpanzee, Pan paniscus) dramatically illustrated his use of this strategy when he was being taught how to create stone tools by a human expert. Unable to replicate the manipulations of the human expert, Kanzi eventually resorted to smashing stones upon others, and simply selected shards with sharp edges, in order to produce his stone tools¹.

[edit] Applications to theorem-proving

Some mathematical proofs can be reliably constructed by working backward. At times, a transformation is required, in order to fix up a solution so that it matches the desired goal.

[edit] Applications to AI

AI programs for medical diagnosis or fault-finding in electrical equipment can use backward-chaining.

[edit] See also

• Abductive reasoning
• George Polya, How to Solve It
• Thinking
• Top-down

[edit] Notes

• Note 1: How Kanzi worked backward from the goal of flint-knapping tools