Chapter 5 . Applications of Adaptive Agents
81
achieved by recurrent evolutionary computation according to Fig. 5.5. embodied
in the agents.
Table 5.1. Predictive rules for annual dynamics of Microcystis and Oscillatoria
evolved from the Lake Kasumigaura database (Bobbin and Recknagel2001)
Evolved Rules fOT Microcystis
IF (TEMP > 29°C) AND
IF (DTP > 74.2 /lgll) AND
IF (pH > 8.15)
TIfEN MICROCYSTIS » 50,000 cells/ml
ELSE MICROCYSTIS < 50,000 cells/ml
5.4.1.2
Adaptive Agents Bank
Evolved Rules fOT Oscillatoria
IF (NH4 < 236 JLgIl) AND
IF (8.01 < pH < 9.37) AND
IF (60 < SECCHI < 103 cm) AND
AND IF (DTP > 22.3 JLgIl)
TIfEN OSCIILA TORIA » 50,000 cells/ml
ELSE OSCIILATORIA< 50,000 cells/ml
Natural ecosystems are characterised by redundancy in their composition and
structure. They gain a certain degree of resilience to changing environmental
conditions depending on the extent of redundancy. In order to develop adaptive
agent models that gain such resilience to environmental changes, they need to
have redundancy in their composition as weIl. Therefore, a bank of alternative and
additional ecological agents for algal species will be developed occurring
seasonally and locally in specific lakes under certain environmental conditions as
reflected in the lake database (Tab. 5.3.). Evolutionary computation (EC) will be
applied according to Fig. 5.5. to develop these algae-specific agents from the lake
database that currently contains multivariate time-series of nine lakes different in
eutrophication, climate and morphology. The range of conditions in the database
will result in alternative agents for the same species/population resting in an agent
bank.
81
achieved by recurrent evolutionary computation according to Fig. 5.5. embodied
in the agents.
Table 5.1. Predictive rules for annual dynamics of Microcystis and Oscillatoria
evolved from the Lake Kasumigaura database (Bobbin and Recknagel2001)
Evolved Rules fOT Microcystis
IF (TEMP > 29°C) AND
IF (DTP > 74.2 /lgll) AND
IF (pH > 8.15)
TIfEN MICROCYSTIS » 50,000 cells/ml
ELSE MICROCYSTIS < 50,000 cells/ml
5.4.1.2
Adaptive Agents Bank
Evolved Rules fOT Oscillatoria
IF (NH4 < 236 JLgIl) AND
IF (8.01 < pH < 9.37) AND
IF (60 < SECCHI < 103 cm) AND
AND IF (DTP > 22.3 JLgIl)
TIfEN OSCIILA TORIA » 50,000 cells/ml
ELSE OSCIILATORIA< 50,000 cells/ml
Natural ecosystems are characterised by redundancy in their composition and
structure. They gain a certain degree of resilience to changing environmental
conditions depending on the extent of redundancy. In order to develop adaptive
agent models that gain such resilience to environmental changes, they need to
have redundancy in their composition as weIl. Therefore, a bank of alternative and
additional ecological agents for algal species will be developed occurring
seasonally and locally in specific lakes under certain environmental conditions as
reflected in the lake database (Tab. 5.3.). Evolutionary computation (EC) will be
applied according to Fig. 5.5. to develop these algae-specific agents from the lake
database that currently contains multivariate time-series of nine lakes different in
eutrophication, climate and morphology. The range of conditions in the database
will result in alternative agents for the same species/population resting in an agent
bank.