Chapter 4
Ecological Applications of Evolutionary
Computation
P.A. Whigham . G.ß. Fogel
4.1
I ntrod uction
Ecological modelling covers a broad range of techniques, concepts and fields of
study. Ecosystems display many complex structures, such as hierarchical
organization, and many components interacting at different temporal and spatial
scales. These components, and the overall structure of any ecosystem, have also
been derived through complex adaptations that produce such features as
speciation, symbiosis and community structures.
The most basic ecological models are concerned with the behavior of a single
species, group of species or community and their variation in time. These models
can be expressed as predictive functions, based on independent variables such as
environmental factors, other species and other abiotic factors. More complex
models capture the coupled interaction between species and may incorporate
spatial extent when describing the system. The most complex models attempt to
capture the spatial and temporal patterns of an ecosystem at several levels of
description. These models aim to understand the biological and evolutionary
mechanisms that have produced an observed ecosystem and to explore the basic
properties that produce diversity and structure in an idealized system.
Evolutionary computation is a discipline that makes use of principals from
natural evolution to evolve solutions to complex computational problems. These
techniques have been applied to such diverse areas as optimization, inductive
modelling, constructing characteristic features of biological systems and as
theoretical models of social and population-based interactions. Their capability to
handle large search spaces (in terms of possible decisions or model constructs) in
an efficient manner, and their similarity to biological systems, such as
genotype/phenotype mapping, structured populations of individuals, adaptation
and evolution makes them ideal candidates for constructing ecological models.
The advent of technologies such as global positioning systems, satellite images,
improved field data collection services and unobtrusive tracking devices have
Ecological Applications of Evolutionary
Computation
P.A. Whigham . G.ß. Fogel
4.1
I ntrod uction
Ecological modelling covers a broad range of techniques, concepts and fields of
study. Ecosystems display many complex structures, such as hierarchical
organization, and many components interacting at different temporal and spatial
scales. These components, and the overall structure of any ecosystem, have also
been derived through complex adaptations that produce such features as
speciation, symbiosis and community structures.
The most basic ecological models are concerned with the behavior of a single
species, group of species or community and their variation in time. These models
can be expressed as predictive functions, based on independent variables such as
environmental factors, other species and other abiotic factors. More complex
models capture the coupled interaction between species and may incorporate
spatial extent when describing the system. The most complex models attempt to
capture the spatial and temporal patterns of an ecosystem at several levels of
description. These models aim to understand the biological and evolutionary
mechanisms that have produced an observed ecosystem and to explore the basic
properties that produce diversity and structure in an idealized system.
Evolutionary computation is a discipline that makes use of principals from
natural evolution to evolve solutions to complex computational problems. These
techniques have been applied to such diverse areas as optimization, inductive
modelling, constructing characteristic features of biological systems and as
theoretical models of social and population-based interactions. Their capability to
handle large search spaces (in terms of possible decisions or model constructs) in
an efficient manner, and their similarity to biological systems, such as
genotype/phenotype mapping, structured populations of individuals, adaptation
and evolution makes them ideal candidates for constructing ecological models.
The advent of technologies such as global positioning systems, satellite images,
improved field data collection services and unobtrusive tracking devices have