ARTIFICIAL INTELLIGENCE IN COMPUTER GAMES

WHAT IS Artificial intelligence?

At the core, Artificial Intelligence(AI) is an idea of the ideal and rational behaviour of a machine. There are eight definitions, which are further categorized into four.

In simple words, when a machine exhibits signs of intelligence, that is when it has the ability to act and think rationally when it is called as an intelligent agent. And this is called artificial intelligence. There are two major aspects in the study of AI, agent and environment. An agent is a machine which has AI in it. And the environment is where an agent dwells. There are two components in an agent, actuator and sensor. A sensor receives a sequence of percepts from the environment. And actuator acts upon the suggestion or advice given by the AI component of the agent. AI component reads the set of precepts detected by the sensor and computes it into a sequence of reasonable actions. these activities are performed by the actuators.

CONFLICT OF INTEREST

There has always been a commotion between robotics and gaming communities. People often argued that game AI is not really an AI. Because game AI depends on predefined and automated computation and has limited responses. when agent actions are predefined then there isn’t any sign of intelligence. but with an addition of learning component in a game bot or NPC we could say that game AI can be considered as AI.

 

NON PLAYER CHARACTER(NPC)

AI has much application in the game industry and one of them is  NPCs moment and actions. In a game, NPC is considered to be one of the essential parts of a game. The AI of an NPC is programmed with automated responses for every variety of situations, some argue that these responses are predefined and the NPCs only try to mimic the human behaviour and shows us a false sense of intellect. But it is only true in a few aspects of AI. For instance, consider a metal gear solid v. It was a revolutionary game of 2015. In this game, the NPCs cannot see the player unless or until he comes within the vision radius. And if the player makes any sound by waking over something the NPC detects the player. And there are so many other aspects that prove development in game AI in recent years.

 

 APPLICATIONS OF AI IN GAME INDUSTRY

Player-experience modelling.

Procedural-content generation.

Data mining on user behaviour.

Alternate approaches to NPCs.

Pathfinder Algorithm

This technique is used on NPCs to move around the game environment. It is a simple algorithm which uses a combination of few search algorithm to find their path around in the game environment.

Mostly it considers all possible paths, compares all the attributes such as cost, time, distance etc. and picks out the best possible path from the source point to the final/goal point.

There are two types of a path-finding algorithm :

A* search algorithm.

Dijkstra algorithm.

D* algorithm

Any-angle path planning algorithm

A* search algorithm

A* is an informed search strategy algorithm which is used in graph traversal and path-finding. It is considered to be one of the best search strategies under proper and admissible situations.

Researchers created a pathfinding algorithm for ‘shakey the robot’, it was used for robots’ navigation, which was then called A1 which led to A* with few minor changes.

A* starts out by finding out the distance from every node in the graph to h the goal node. Using this information it finds out a best possible path from the start node to the goal node.

A* selects the path that has least f(n)

f(n)=h(n)+g(n)

Where h(n) is heuristic, it estimates the cheapest path from n to goal node, g(n) is the cost of a path travelled till the node n.

D* algorithm

D* comes under-informed incremental search strategy.

The algorithm works assuming that there are no obstacles in the path and finds out the shortest path from the current node to a goal node.

Then the bot follows the path and it might encounter obstacles or new map information, then it will change its path accordingly.

More than often it has to re-plan the shortest path and it has to be fast,   hence it uses incremental search algorithms’ experience with previous problems, assuming that it has similar coordinates to the goal node.

Any angle path planning

This algorithm is used to search for the path between two points in a space and allows the bot to take turns in any particular angle

Whenever a blocked cell is encountered this algorithm knows which side to traverse, so this is better than both A* and grid constraint algorithms.

 

 Artificial Intelligence PARANOIA

Artificial Intelligence, in general, is used for assistance and to help humans in all possible ways. AI overthrowing humans or forming an insurgency to exterminate humans is just fictional and moreover paranoia. With everything that AI does, people often assume that the AI agent has emotions. AI in present day neither has IQ nor EQ. Even though in future if AI is made to exhibit IQ, there is a negligible chance of an intelligent agent acquiring EQ.

 

 

Leave a Reply

Your email address will not be published. Required fields are marked *