Axiomatic Approaches to Bargaining

Download Report

Transcript Axiomatic Approaches to Bargaining

5. Alternative Approaches
1. Introduction
2. Individual Decision Making
3. Basic Topics in Game Theory
4. The Theories and the Real World
5. Alternative Approaches
5.1 Psychology
5.2 Evolution
5.3 Artificial Intelligence
Strategic Bahavior in Business and
Econ
Computer Science tools for the Social Sciences
Computer Science contains several tools and techniques
that are useful to the Social Sciences. Among others we
have:
Theoretical artifacts that represent “computer
behavior” (Artificial Intelligence). Some can be adapted
to represent “human behavior”
Computational devices and techniques (soft and
hardware) that can be used to simulate human behavior
Strategic Bahavior in Business and
Econ
Theoretical artifacts
Consist of theoretical models of behavior designed
and developed to represent and to study the functioning
of computers. They are the basic tools of AI:
Cellular Automata
Turing Machines
Neural Networks
Perceptrons
Finite Automata
···
Strategic Bahavior in Business and
Econ
Finite Automata
A Finite Automaton (Automaton for short) is a
“device” that consists of:
Several internal states that represent the different “moods” of the
automaton
Ex: “happy”, “sad”
Several possible outputs that represent the different actions that
the automaton can take at each state
Ex: “give flowers”, “give nothing”
Several possible inputs that represent the different information that
the automaton can receive from the environment at each state
Ex: “a smile”, “nothing”
Several transitions that represent the change from one state to
another depending on the input obtained from the environment
Strategic Bahavior in Business and
Econ
Example: The “regular” Lion
In the presence of another lion, act friendly unless
it is behaves aggressively. In such case you should
act aggressively as well
In the presence of a zebra, chase it and share the
meal with the other lions and cabs
States: “Relaxed”, “Aggressive”
Actions: “Friendly”, “Attack”
Inputs: “Friendly Lion”, “Aggressive Lion”
“Friendly zebra”, “Aggressive zebra”
Transitions: (in the graph)
Strategic Bahavior in Business and
Econ
Example: The “regular” Lion
Friendly Lion
Aggressive Lion
Friendly
Aggressive zebra
Attack
Friendly zebra
Aggressive
Relaxed
Strategic Bahavior in Business and
Econ
Example: The “crazy” Lion
In the presence of another lion, chase it and share
the meal with the other lions and cabs
In the presence of a zebra, act friendly unless
it is behaves aggressively. In such case you should
act aggressively as well
Strategic Bahavior in Business and
Econ
Example: The “crazy” Lion
Friendly zebra
Aggressive zebra
Friendly
Aggressive Lion
Attack
Friendly Lion
Relaxed
Aggressive
Strategic Bahavior in Business and
Econ
Example: The “friendly” Lion
In the presence of another lion, act friendly unless
it is behaves aggressively. In such case you should
act aggressively as well
In the presence of a zebra, act friendly unless
it is behaves aggressively. In such case you should
act aggressively as well
Strategic Bahavior in Business and
Econ
Example: The “friendly” Lion
Friendly zebra
Aggressive zebra
Friendly
Aggressive Lion
Attack
Friendly Lion
Relaxed
Aggressive
Strategic Bahavior in Business and
Econ
Example: The Repeated Prisoners' Dilemma
Strategic Bahavior in Business and
Econ
Example: The Repeated Prisoners' Dilemma
Strategic Bahavior in Business and
Econ
The contributions of automata
Controlling the different elements of the automaton
we can represent different degrees of rationality
(larger memory, different information processing power,
different number of possible actions, etc)
Fixing the size of the automata available to the players
of a game, we control the degree of complexity of the
strategies they can choose
Automata, though, does not control for the
computational ability
The main results from “fully rational” game theory do
not change much if we consider “boundedly rational”
game theory with automata representation
Strategic Bahavior in Business and
Econ
Computer Science tools for the Social Sciences
Computer Science contains several tools and techniques
that are useful to the Social Sciences. Among others we
have:
Theoretical constructs that represent “computer
behavior” (Artificial Intelligence). Some can be adapted
to represent “human behavior”
Computational devices and techniques (soft and
hardware) that can be used to simulate human behavior
Strategic Bahavior in Business and
Econ
Computer Simulations
Using the power of modern computers we can
simulate the behavior of economic agents with
a high degree or realism.
This way we do not need to make strong
simplifying assumptions as in the formal
mathematical models
Strategic Bahavior in Business and
Econ
Strategic Bahavior in Business and
Econ
Strategic Bahavior in Business and
Econ
Examples:
The Schellling's Segregation Model
The Hotelling's Spatial Competition Model
Strategic Bahavior in Business and
Econ
The First Simulation in Economics
Strategic Bahavior in Business and
Econ
The First Simulation in Economics
Strategic Bahavior in Business and
Econ
Evolution and Artificial Intelligence
We can combine:
The theoretical artifacts from AI (automata)
The ideas from evolutionary theory
The computational techniques (simulations)
To control for:
The complexity of the strategies and the information
processing
The computational abilities
The need for simplifying assumptions
Strategic Bahavior in Business and
Econ