40
X-Machines for Agent-Based Modeling: FLAME Perspectives
“Why can’t we build, once and for all, machines that grow and improve themselves by learning from experience? Why can’t we simply explain what we want, and then let our machines do experiments or read
some books or go to school, the sort of things people do. Our machines
today do no such things.” [135]
Figure 2.2 depicts different research, stemming from the umbrella of artificial intelligence. Each method is designed for specific purposes, such as genetic
algorithms are efficient optimization techniques to search in NP-hard problems
or neural networks used to encourage speech and voice recognition in software
and other areas. Researchers [167] compared the efficiency of these techniques,
when applied to similar problems and drawn conclusions for computational
efficiency, resources and time.
Advances in parallel computers and architectures have aided research
in multiple areas of science and engineering, with ABM platforms working with researchers with less programming experience. Sante Fe has produced various agent-based models of various kinds like artificial stock market,
molecular structures and more. Details can be found at the main website
(http://www.santefe.edu/).
2.7 Influence of Other Research Areas on ABM
Markov modeling using Markov decision processes. These models are
based on mathematical expressions of Turing machine models. The algorithm involves executing a number of rules, encoded on a symbol string.
Markov models can be expressed as chains containing stochastic processes whose states change with time. These state changes carry conditional probabilities associated with them. The future states are independent of past states.
Markov decision processes use a reward function attached with Markov
chains. For every transition, the state receives a reward that affects
the transition probability of the state. Using reinforcement learning,
these systems are useful in dynamic programming problems and training
problems such as using unobservable states in hidden Markov models.
Neural networks. These are recreate biological structures of the neuron activity in organisms. The various nodes are connected to each other, with
each connection carrying weights for the path to process data. Neural
networks can be trained using real data. The simulated data can then
be verified if it produces similar results.
X-Machines for Agent-Based Modeling: FLAME Perspectives
“Why can’t we build, once and for all, machines that grow and improve themselves by learning from experience? Why can’t we simply explain what we want, and then let our machines do experiments or read
some books or go to school, the sort of things people do. Our machines
today do no such things.” [135]
Figure 2.2 depicts different research, stemming from the umbrella of artificial intelligence. Each method is designed for specific purposes, such as genetic
algorithms are efficient optimization techniques to search in NP-hard problems
or neural networks used to encourage speech and voice recognition in software
and other areas. Researchers [167] compared the efficiency of these techniques,
when applied to similar problems and drawn conclusions for computational
efficiency, resources and time.
Advances in parallel computers and architectures have aided research
in multiple areas of science and engineering, with ABM platforms working with researchers with less programming experience. Sante Fe has produced various agent-based models of various kinds like artificial stock market,
molecular structures and more. Details can be found at the main website
(http://www.santefe.edu/).
2.7 Influence of Other Research Areas on ABM
Markov modeling using Markov decision processes. These models are
based on mathematical expressions of Turing machine models. The algorithm involves executing a number of rules, encoded on a symbol string.
Markov models can be expressed as chains containing stochastic processes whose states change with time. These state changes carry conditional probabilities associated with them. The future states are independent of past states.
Markov decision processes use a reward function attached with Markov
chains. For every transition, the state receives a reward that affects
the transition probability of the state. Using reinforcement learning,
these systems are useful in dynamic programming problems and training
problems such as using unobservable states in hidden Markov models.
Neural networks. These are recreate biological structures of the neuron activity in organisms. The various nodes are connected to each other, with
each connection carrying weights for the path to process data. Neural
networks can be trained using real data. The simulated data can then
be verified if it produces similar results.