Network behaviours

Network Behaviours

Behavioural classes

There are 10 classes that determine the behaviour of a network node within a running network.

Of these 10, there are 4 concrete classes which already provide sufficient functionality for their purpose, but may still be extended.

TODO: Add final descriptions of attributes and functions to the following, as well as consider the changes to genetic sequence

[InputProcess]
[OutputProcess]
[StorageProcess]
[NodeProperties]

In addition to these, there are also 6 abstract classes which must be extended. Note that some do not necessarily need to be extended to create a functioning network.

TODO: Add final descriptions of attributes and functions to the following

[FiringCondition]
[TargetSelection]
[TransmissionContent]
[DataProcessing]
[EnergyEconomics]
[LifeCycle]

Concurrency verification

Because this system, and in particular these behavioural classes, employ a high degree of concurrency, it is necessary to do a formal concurrent verification. This has been done for the above network node behaviour with the LTSATool. This modelled the core behaviour of the network nodes, as well as some extensions. The lts file is located in the verification folder in the source code.

Focused behavioural models

These are extensions of the behavioural classes above that only encompass a few or even a single behavioural domain. These, for the most part, do not constitute a complete and operational behavioural model on their own. They must be combined with other models, and if required to actually run, a set of network initial conditions.

The idea of these sectons is to eventually produce implementations of just one or two behavioural classes for each model, depending on how many they need to use, which may be easily combined with other implementations to quickly produces larger and more complex behavioural models.

TODO: Descriptions of each of the following

Simple testing models:
[Fast random behavioural model] - Uses the [FastFiringCondition] class

Signal routing focused models (TargetSelection focused):
[Simple synapse strengthening model]
[Persistent signal model]
[Stimuli dissemination model]
[3D mapping model]

Configuration manipulation models (NodeProperties focused):
[GeneticAlgorithmNodeProperties] extension of [NodeProperties]

Network growth and decay models (LifeCycle focused):
[TrafficBasedLifeCycle] extension of [LifeCycle]

Broad behavioural models

These are compositions of the previous behavioural models combined with additional specific functionality and initial network configurations, designed to provoke more interesting and less homogeneous simulations than a single set of the above models would on their own.

[Model maker composition]
[Ecological model]