Forward-chaining and Backward-chaining

1. Forward-Chaining Algorithm

Definition:

Forward-Chaining is a data-driven inference method.
It starts from known facts and applies rules to infer new facts until the goal is reached (or no more inferences can be made).

How it works:

1. Start with a set of facts (knowledge base).

2. Apply all rules whose preconditions match the current facts.

3. Add the rule’s conclusions (new facts) to the knowledge base.

4. Repeat until:

   • The goal is inferred, or

   • No more rules can be applied.

Example:

Facts:

• Sun is shining → Sunny

• Sunny → Go outside

• Go outside → Happy

Initial Fact:

• Sun is shining

Goal:

• Happy

Inference:

1. Sun is shining → infer Sunny

2. Sunny → infer Go outside

3. Go outside → infer Happy

Goal is reached.

Advantages:

• Suitable when all data is known upfront.

• Works well in real-time systems, such as expert systems and diagnostics.

---

2. Backward-Chaining Algorithm

Definition:

Backward-Chaining is a goal-driven inference method.
It starts with the goal and works backward to determine what facts must be true to satisfy that goal.

How it works:

1. Start with a goal.

2. Look for rules that can produce that goal in the conclusion.

3. Make the preconditions of that rule the new subgoals.

4. Repeat the process until you reach known facts or fail.

Example:

Rules:

• Sunny → Go outside

• Go outside → Happy

• Sun is shining → Sunny

Goal:

• Happy

Inference:

1. To prove Happy, need Go outside

2. To prove Go outside, need Sunny

3. To prove Sunny, need Sun is shining

If Sun is shining is known, then the goal Happy is provable.

Advantages:

• Efficient when the goal is known, and you want to verify if it can be satisfied.

• Used in logic programming (e.g., Prolog), theorem proving, and question answering systems.

---

Forward vs Backward Chaining

Feature	Forward-Chaining	Backward-Chaining
Approach	Data-driven	Goal-driven
Starts from	Known facts	Desired conclusion
Stops when	Goal is derived	Facts supporting goal are found
Use Case	Expert systems, diagnosis	Logic programming, proof systems
Direction	Bottom-up inference	Top-down reasoning

---

Summary

• Forward-Chaining: Pushes known facts forward to derive conclusions.

• Backward-Chaining: Pulls from the goal backward to verify if facts can support it.

Both are fundamental to AI reasoning, rule-based systems, and knowledge inference.