202
X-Machines for Agent-Based Modeling: FLAME Perspectives
the list is acted upon in the same manner. In agents, the messages in
the sphere of influence may vary allowing agents to display different behaviors depending on where they are located. This is particularly useful
when modeling realistic biological models.
• Heterogeneous populations. Different agents who differ in memory can
be introduced together in the same simulation. This can produce more
interesting results as it brings heterogeneity and how agent internal characteristics can influence results. This cannot be done in equation models
as these assume a homogeneous population.
• Bounded rationality. Agents will act depending on their surroundings
producing emergent phenomena. This cannot be programmed in an
equation model.
• Different scenarios during simulations. Easily different conditions can be
introduced to test the model across various conditions. This would not
require doing any changes to the agent-based models. Simulations can
be stopped halfway, conditions can be changed or new agents can be
introduced at adhoc and then simulations can be preceded.
• Large amount of data produced. This is a problematic task to analyze
large amounts of data being produced by agent-based models as compared to equation models. Sometimes it is good to find patterns which
may not have been thought of previously, but this can be a cumbersome
task and may require additional intelligent data mining algorithms at a
later stage.
However, it largely depends on the research questions being investigated when
a model is being written. In all cases there is a learning curve for biologists
and computer scientists to understand which to use and why.
7.5 Modeling Pharaoh’s Ants (Monomorium pharaonis)
Pharaoh’s ant is a 2-mm monomorphic pest ant forming colonies with
less than 2500 workers. These ants have poor vision, making them wholly
reliant on the pheromones deposited for path directions. Unlike any other ant
species, they deposit trail pheromones constitutively when outside the nest,
forming branching networks of pheromone trails even before the food sources
are discovered [23].
In the model, ant agents are characterized by their identity number, nutritional status, current direction and environmental locations. Each ant agent
X-Machines for Agent-Based Modeling: FLAME Perspectives
the list is acted upon in the same manner. In agents, the messages in
the sphere of influence may vary allowing agents to display different behaviors depending on where they are located. This is particularly useful
when modeling realistic biological models.
• Heterogeneous populations. Different agents who differ in memory can
be introduced together in the same simulation. This can produce more
interesting results as it brings heterogeneity and how agent internal characteristics can influence results. This cannot be done in equation models
as these assume a homogeneous population.
• Bounded rationality. Agents will act depending on their surroundings
producing emergent phenomena. This cannot be programmed in an
equation model.
• Different scenarios during simulations. Easily different conditions can be
introduced to test the model across various conditions. This would not
require doing any changes to the agent-based models. Simulations can
be stopped halfway, conditions can be changed or new agents can be
introduced at adhoc and then simulations can be preceded.
• Large amount of data produced. This is a problematic task to analyze
large amounts of data being produced by agent-based models as compared to equation models. Sometimes it is good to find patterns which
may not have been thought of previously, but this can be a cumbersome
task and may require additional intelligent data mining algorithms at a
later stage.
However, it largely depends on the research questions being investigated when
a model is being written. In all cases there is a learning curve for biologists
and computer scientists to understand which to use and why.
7.5 Modeling Pharaoh’s Ants (Monomorium pharaonis)
Pharaoh’s ant is a 2-mm monomorphic pest ant forming colonies with
less than 2500 workers. These ants have poor vision, making them wholly
reliant on the pheromones deposited for path directions. Unlike any other ant
species, they deposit trail pheromones constitutively when outside the nest,
forming branching networks of pheromone trails even before the food sources
are discovered [23].
In the model, ant agents are characterized by their identity number, nutritional status, current direction and environmental locations. Each ant agent