88
X-Machines for Agent-Based Modeling: FLAME Perspectives
ulations begin with a rigourous set of assumptions and generate data which
are analyzed via induction.
Tesfatsion and Axelrod [194] discussed specific goals for agent-based researchers to pursue.
Empirical. How will particular large-scale phenomena perform such as emergence of social norms? These are analyzed to see if they pursue with
global irregularities, to find why certain behaviors persist.
Normative. How are agent-based models used to discover good designs? Investigate key issues like efficiency and order.
Heuristic. Can complex behaviors be attained via simple interactions? It is
difficult to predict behavior of large-scale systems with interactions on
smaller scales.
Methodological Advancement. Study models and simulation data are
produced.
Conway’s Game of Life is one of the earliest examples using cellular automata grids, to display a group of cells interacting with each other [72].
Using only four simple rules, through simulation, the game could present new
patterns depending on neighboring cells. For every cell, the four rules are
1. If current cell is alive and has less than two neighboring cells - die due
to lack of social activity.
2. If current cell is alive and has more than three neighboring cells - die
due to overcrowding.
3. If current cell has two or three live neighbors - survive to next time step.
4. If current cell is a dead cell but has three live neighbors - become alive.
.
Since this example, computer simulations have taken a long journey to
more complicated computer programs using agent-based modeling and parallel
computing. Figure 5.1 shows a snapshot of a game of life run. The black dots
represent live cells. All cells follow the four simple rules, producing the various
patterns.
Schelling [169] used a segregation model to show that, by adding a small
preference factor, societies can emerge into segregated ones. The model, based
on 2D grid landscape, had individuals represented by different colors. Every
colored individual would check for the following rule:
If more than 33% of adjacent individuals or cells were of different
color, the cell should randomly move to new position.
X-Machines for Agent-Based Modeling: FLAME Perspectives
ulations begin with a rigourous set of assumptions and generate data which
are analyzed via induction.
Tesfatsion and Axelrod [194] discussed specific goals for agent-based researchers to pursue.
Empirical. How will particular large-scale phenomena perform such as emergence of social norms? These are analyzed to see if they pursue with
global irregularities, to find why certain behaviors persist.
Normative. How are agent-based models used to discover good designs? Investigate key issues like efficiency and order.
Heuristic. Can complex behaviors be attained via simple interactions? It is
difficult to predict behavior of large-scale systems with interactions on
smaller scales.
Methodological Advancement. Study models and simulation data are
produced.
Conway’s Game of Life is one of the earliest examples using cellular automata grids, to display a group of cells interacting with each other [72].
Using only four simple rules, through simulation, the game could present new
patterns depending on neighboring cells. For every cell, the four rules are
1. If current cell is alive and has less than two neighboring cells - die due
to lack of social activity.
2. If current cell is alive and has more than three neighboring cells - die
due to overcrowding.
3. If current cell has two or three live neighbors - survive to next time step.
4. If current cell is a dead cell but has three live neighbors - become alive.
.
Since this example, computer simulations have taken a long journey to
more complicated computer programs using agent-based modeling and parallel
computing. Figure 5.1 shows a snapshot of a game of life run. The black dots
represent live cells. All cells follow the four simple rules, producing the various
patterns.
Schelling [169] used a segregation model to show that, by adding a small
preference factor, societies can emerge into segregated ones. The model, based
on 2D grid landscape, had individuals represented by different colors. Every
colored individual would check for the following rule:
If more than 33% of adjacent individuals or cells were of different
color, the cell should randomly move to new position.