Transcript SurvivalRobots

Survival Robots
Karl Lieberherr
• To teach development of reliable software and to
experimentally research combinatorial optimization.
• Offer and buy food (both life-energy increasing and
decreasing) in an artificial world.
• Food = financial derivatives. Raw materials and finished
products to process the food.
• Challenge: Hide and find secrets.
• Unreliable robots die due to bad trading decisions and
world rule violations.
• Summary: Survival Robots help to transmit CS
knowledge and to create new CS knowledge about
optimization.
Survival Robots
Karl Lieberherr
• Survival means: good trading decisions / follow
market rules: reliable software.
• To teach development of reliable software and to
experimentally research artificial markets of
relevance to computer science (P=NP problem).
• Offer and buy derivatives, raw materials, finished
products. Challenge: Hide and find secrets.
• Summary: Survival Robots help to transmit CS
knowledge and help to create new CS
knowledge: Functional Adaptive Programming,
Experimental Algorithmics, New Algorithms for
NP problems, Reliable Software.
Artificial Markets Driving
CS Innovation
Karl Lieberherr
Computer Science Knowledge
produce reliable software
learn
Artificial Markets
(derivatives,
raw materials,
finished products)
Trading Robots
survive?
Real world problems we solve: Producing reliable and evolutionary software
(through Functional Adaptive Programming),
Improving combinatorial optimization algorithms.
Scientific Contributions: Programming technology with provable properties,
Theory of Computation, Study of artificial markets.
Artificial Markets Driving
CS Innovation / Education
Karl Lieberherr
Computer Science Knowledge X, Design Artificial Market for X
produce reliable software
learn
Run Simulation of
Artificial Market
Trading Robots
survive?
Real world problems we solve: Producing reliable and evolutionary software
(through Functional Adaptive Programming),
Scientific Contributions: Programming technology with provable properties,
Study of artificial markets.
Survival Robots
Karl Lieberherr
Computer Science Knowledge
produce reliable software
learn
Artificial Markets
(derivatives,
raw materials,
finished products)
Survival Robots
survive?
Real world problems we solve: Producing reliable and evolutionary software
(through Functional Adaptive Programming),
Improving combinatorial optimization algorithms.
Survival Robots
Karl Lieberherr
Computer Science Knowledge
produce reliable software
learn
Artificial Markets
(derivatives,
raw materials,
finished products)
Survival Robots
survive?
Scientific Contributions: Programming technology with provable properties,
Theory of Computation, Study of artificial markets.
Survival Robots
Karl Lieberherr
Computer Science Knowledge
produce reliable software
learn
Survival Robots
Artificial Markets
survive?
Artificial Markets: (derivatives,
raw materials,
finished products)
CSU 670
Artificial Market Driving
Software Development + X
Education
Computer Science Knowledge X, Design Artificial Market for X
produce reliable software
learn
Run Simulation of
Artificial Market
Trading Robots
survive?
X = CSP
Learn about Software Development: Agile Software Development,
Integration testing, security policies, style rules (Law of Demeter,
Terminal Buffer Rule), reliable evolution using DemeterF, etc..