50
X-Machines for Agent-Based Modeling: FLAME Perspectives
FIGURE 3.5: Agile agent development process.
9. For simplicity, try not to increase work if not needed.
10. Allow products to emerge from teams.
11. Teams continuously reflect their performance, share ideas and learn,
becoming effective.
Developing agent models, using agile allows multiple domain experts to
work together, to develop software models. It enables computer scientists to
work closely with domain experts, to build a model based on domain requirements. Testing of model and verification is also continuously done at every
stage of the release, minimizing risk of wrong assumptions being implemented
in the model.
Agent-based models are difficult to implement due to sheer complexity of
models. Through the process in Figure 3.5, domain experts can interact closely
with modelers, to monitor model development and research hypotheses. However, with these advantages, the process sometimes slows down development
and introduces the need for continuous client involvement. But at the end of
every cycle, as the model matures, the clients are able to monitor and develop
ideas and test these through their models before releasing them to the research
domain.
X-Machines for Agent-Based Modeling: FLAME Perspectives
FIGURE 3.5: Agile agent development process.
9. For simplicity, try not to increase work if not needed.
10. Allow products to emerge from teams.
11. Teams continuously reflect their performance, share ideas and learn,
becoming effective.
Developing agent models, using agile allows multiple domain experts to
work together, to develop software models. It enables computer scientists to
work closely with domain experts, to build a model based on domain requirements. Testing of model and verification is also continuously done at every
stage of the release, minimizing risk of wrong assumptions being implemented
in the model.
Agent-based models are difficult to implement due to sheer complexity of
models. Through the process in Figure 3.5, domain experts can interact closely
with modelers, to monitor model development and research hypotheses. However, with these advantages, the process sometimes slows down development
and introduces the need for continuous client involvement. But at the end of
every cycle, as the model matures, the clients are able to monitor and develop
ideas and test these through their models before releasing them to the research
domain.