INTRODUCTION
Artificial Intelligence (AI) is a subject that often appears complex when students encounter it for the first time. The syllabus is extensive, the terminology is unfamiliar, and many algorithms seem abstract. During classroom interactions, students frequently ask where these concepts are actually applied. AI, however, is not a single tool or technology. It is a structured way of understanding how machines can solve problems, make decisions, and learn from experience.
Artificial Intelligence focuses on building systems that can perform tasks requiring intelligence when done by humans. These tasks include reasoning, problem solving, learning, and decision-making. Instead of emotions or consciousness, AI emphasizes logic, structure, and adaptability.
Basically Machines which behaves like humans and work like humans can be considered as Artificial Intelligence.They will be good in executing these parameters like Reasoning,Learning, problem solving and Perception.
AGENTS/INTELLIGENT AGENTS
At the core of AI lies the concept of intelligent agents. An agent is an entity that perceives its environment and takes actions accordingly. Agents can be simple or complex depending on how much information they use. Some agents react directly to inputs, while others plan actions based on goals or learned experience.
Agents Program first percept/sense from the environment using current /history information and then try to make the decision accordingly and finally via actuators try to finally execute/implement that action.
GOALS OF AGENT
There are 3 major goals of agents as below:
- High Performance
- Optimized Result
- Rational Action (Right)
Important Factors to design any Agent
There are generally 4 factors like Performance, Environment, Actions and Sensors. One real time application which we can relate to it is Automated cars
Types Of Agents
There are different types of agents which are as below:
1)Simple Reflex Agents
- Act only on the basis of current perception.
- Ignore the rest of percept history
- Works on based on IF-THEN Rules
- Environment should be fully observable
- Fully observable –Agent has complete knowledge about the environment.eg->Chess, tic-tac-toe game
- Partially Observable-Agent has partial knowledge about the environment.Eg->In a automated self driving car on the road ,agent doesn’t have information about how many other cars are riding simultaneously,speed of each car etc.
2) Model Based Reflex Agent
- Partially observable environment
- Store percept history
- E.g->Waymo(Google Project)Self Driving car in Arizona(USA)
3) Goal Based Agents
- Expansion of Model based reflex agents
- Desirable Situation
- Searching and Planning
- Eg->For Planning a trip to ladakh ,we would search for shortest route ,safe path and plan before to reach the destination
4) UTILITY BASED AGENTS
- Focus on utility not goal
- Partially Observable environment
- Deals with happy and unhappy state
- Eg->From the environment the perception happens via sensorand then decision and action is taken accordingly.
PROBLEM SOLVING
Problem solving is an important application of Artificial Intelligence.Search is one of the most fundamental problem-solving techniques in AI. Many problems can be represented as a search through possible states. Uninformed search methods such as Breadth First Search (BFS) and Depth First Search (DFS) explore the problem space without additional knowledge. Bidirectional search improves efficiency by searching from both the start and goal states.
- UNINFORMED SEARCHING
- Search without information
- No Knowledge
- Time consuming
- More Complexity with respect to time space
- Optimized/optimum solution
- E.g->DFS,BFS
- INFORMED SEARCHING
- Search With information
- Use knowledge to find steps to solutions
- Quick Solutions
- Less Complexity with respect to time space
- Good Solution
- E.g-?A star,Best First Search,Heuristic DFS
BFS (BLIND FIRST SEARCH)
- It Provides Uninformed search technique
- Also called brute force or blind search
- Also called level search
- Path follows level by level to reach goal state
- Shallowest Node (means nearest node,level by level goes on to each node)
- It Gives complete and optimal solution
- Time complexity is o(b^d),Here b is branch factor and d is depth
DFS (DEPTH FIRST SEARCH)
- Uninformed Search Technique
- It follows Stack(LIFO)
- Deepest Node
- Incomplete /Non Optimal Solution
- Time complexity is o(b^d)
BIDIRECTIONAL SEARCH
- Two simultaneous search from an initial node to goal node and backward from goal node to initial node ,stopping when they meet.
- Here time complexity is o(b^d/2).In worst case it is o(b^d)
- Here space complexity is same but time is less than BFS and DFS
8 PUZZLE PROBLEM
The 8-puzzle problem is commonly used to demonstrate search techniques. Without heuristics, the system explores many unnecessary states. When heuristics are introduced, the search becomes more efficient. The A* algorithm combines the cost so far and an estimated cost to reach the goal, making it one of the most effective informed search methods.
8 PUZZLE PROBLEM WITHOUT HEURISTIC
- It Follows Blind Search(Uninformed)
- BFS
- Time space complexity-o(b^d)
- It has 4 moves (up,down,left,right)
HEURISTIC IN AI
- It is a technique designed to solve a problem quickly
- Used for finding quick solutions
- It reduces time
- Will get good solutions not optimum solutions
- Basically it solve NP Problem in Polynomial time
BEST FIRST SEARCH(INFORMED ,HEURISTIC)
- This follows Priority Quene
- It provides Greedy Method based on least heuristic value.
- It provides good solution
- It provides optimal solution
- Here if node is a GOAL then Return the path from initial to goal node else Generate all successors of node and put the newly generated node onto open according to their heuristic values.
Discussion Questions
How does viewing Artificial Intelligence as a collection of problem-solving approaches change the way we understand individual algorithms such as search or learning methods?
Why is it important to study both informed and uninformed search techniques before moving to advanced learning-based models?
In what ways do concepts like game trees and decision-making algorithms reflect real-world strategic thinking?
How do learning paradigms in AI mirror human learning processes, and where do they fundamentally differ?
Course Relevance
This topic helps students connect abstract AI theories with practical reasoning patterns used in real systems.
It supports BCA and MCA learners in understanding how logical thinking, decision-making, and learning mechanisms are implemented computationally.
The content aligns well with core AI curriculum goals by emphasizing understanding over memorization of algorithms.
Academic Concepts Covered
Intelligent agents and their role in AI systems
• Problem-solving through search and heuristic techniques
• Game playing, decision trees, and optimization strategies
• Constraint satisfaction, knowledge representation, and reasoning
• Supervised, unsupervised, and reinforcement learning models
Teaching Note
– Begin with real-life problem-solving examples before introducing formal algorithms.
-Encourage students to visualize search spaces and decision trees using simple illustrations.
– Focus on conceptual clarity and reasoning flow rather than mathematical complexity.
Learning Objectives
• Explain core Artificial Intelligence concepts using relatable examples.
• Identify the role of search, reasoning, and learning in intelligent systems.
• Relate AI algorithms to topics studied in BCA and MCA curricula.
