Please enable JavaScript.
Coggle requires JavaScript to display documents.
Adversarial search and Multiagent environments - Coggle Diagram
Adversarial search and Multiagent environments
Introduction
When several agents are fighting for the same objective in the same setting.
For example, two vacuum cleaners in the same room, one of which starts searching for dust before the other.
Strategy
The significance of decision-making and planning.
Strategies for the long and short terms.
Both proactive and reactive methods.
AND - OR search trees
The search space for resolving an issue is the sole thing specified by a and-or search tree. It is possible to search the area using various search techniques. This search strategy is predicated on the OR and AND criteria. The agent must select one action out of two options first. Here, the OR condition will be used, and the agent will need to take another action after completing that one. The AND condition will be in effect at that point.
For example, Agent Vacuum has two options: proceed correctly or poorly. The OR situation is this. Agents must complete an additional activity that is related to their prior action after completing an action. Suck dust and go right.
1 action specified at each OR
Includes every outcome at each AND
Goal node at every leaf
Adversarial search
Adversarial search is a process that looks at the issue that occurs when we try to plan ahead of time and discover that other agents are intending to do the opposite.
Game theory
The study of mathematical models of conflict and cooperation between intelligent, rational decision-makers.
Optimizing agent performance.
Strategic decision making.
Interactive decision theory.
Multi - agent environment
Cooperation and coordination.
Talking and bargaining.
Interactions involving several substances.
Zero - sum games
Example: Tictactoe, chess, checkers, othello, go
Abstract
Useful to study
Utility values equal and opposite
States easy to represent
Perfect information
Agents restricted to limited actions and precise rules
Compare with physical games
Modifications for a digital setting.
AI versus human gamers.
Both parallels and divergences.
Solving games
Some decision must be made
Efficiency
Calculating optimal decision infeasible
PRUNING the tree
Challenging
Min Max algorithm
Adversarial search - Max has to consider Min’s moves
It’s just like AND-OR search tree
Normal search - actions lead to goal
Search is a complete depth first that backs up
What happens in the real world
Marketplaces and sectors that are competitive.
Bargaining and negotiating.
Adversarial circumstances and judgment calls.
Prisoner's dilemma
In game theory, the most famous example is the Prisoner's Dilemma. Take the case of two offenders who were apprehended for their crimes. In order to convict them, prosecutors lack concrete proof. But in order to get a confession, authorities take the detainees out of their isolation cells and question them individually in other rooms.
Web links
https://www.econlib.org/library/Enc/GameTheory.html#:~:text=Game%20theory%20is%20the,and%20from%20tennis%20to%20takeovers
https://relevanceai.com/learn/what-is-a-multi-agent-system#:~:text=A%20multi%2Dagent%20system%20(MAS,decisions%20and%20act%20upon%20them
.
https://www.javatpoint.com/ai-adversarial-search
https://www.javatpoint.com/mini-max-algorithm-in-ai
https://www.oreilly.com/library/view/ai-for-game/0596005555/ch04.html
https://github.com/rystrauss/boids?tab=readme-ov-file
Emergence
People adhere to basic laws.
Emergence of complexity.
Behavioral patterns.
Phenomenon at the macro level.
Flocking behavior
Simulated behavior of a group of animals.
Shoaling and education.
Pack hunting and herding.
Craig Reynolds - boids
Created in 1986 by Craig Reynolds, Boids is an artificial intelligence living software that mimics bird swarming behavior.
Rules
Separation - avoid crowding
Alignment - head in average direction
Cohesion - steer towards average position
Steering behavior
Agents' local guiding guidelines.
Avoiding pitfalls.
Matching of velocity.
Centering the flock.
Strandbeest
Living artworks with biological inspiration.
Motions and mechanisms.
Self-proliferation and flexibility.