Problem characteristics in AI refers to finding an optimal way and a good solution to characterize the problem. In addition, problem solving is one of the key concerns in AI. Certainly, we know AI is a vast field and branch of computer science, there can be complex problems to solve.
Let us take a look at some of the major problem characteristics.
Major Problem characteristics
Firstly, we need to know if the problem is decomposable or not. For instance, block world problem or Tower of Hanoi are easily decomposable. In Tower of Hanoi some rules are define to move the disks from source to destination. Moreover, each movement is divided into steps.
Secondly, problem is categorize into following:
Ignorable problems are those which can be solved using simple control structure. For instance, mathematical problems.
Recoverable problems are those where we can use backtracking to solve the problem. For instance 8-puzzle problem.
Irrecoverable problems are those where backtracking is not possible. For instance, chess.
Thirdly, we should know if the problem is universally predictable. Lets take look.
- Find out is it a good solution to the problem. It is categorize into two paths : Any path and Best path. Moreover, water jug problem, 8 puzzle problem comes under any path whereas Travelling Salesman Problem comes under best path.
- Is the solution a state of the world or not. This means is there a consistent information.
- How much knowledge is available for solving the problem and does it require interaction with a person.
Problem characteristics Steps
- Define the problem precisely.
- Give the initial input required.
- Apply knowledge
- Choose best optimal technique for problem solving.
Moreover, some problems are also solved using state space representation. State Space is the set of all reachable state from initial states.
In conclusion, we have learnt about the major problem characteristics in Artificial Intelligence. We have seen how to classify the problem and major steps involved in problem solving.