The introduction and the agent-oriented development guide together present a picture of the kinds of solution an agent-based approach makes possible, and the types of environment they are most suited for. In short, this is where:
- the environment is decentralised,
- involves multiple stakeholders, and
- is inhabited by AEAs representing the different stakeholders who:
- interact autonomously, and
- communicate with one another directly via a peer-to-peer network.
In light of those discussions, on this page we identify a number of application areas for AEA-based solutions. This list is by no means comprehensive. In fact, we are most excited about applications which we have not thought of before.
- Inhabitants: agents representing objects in the IoT (Internet of Things) space. For examples, AEAs paired with real world hardware devices such as drones, laptops, heat sensors, etc. An example is a thermometer agent .
- Interfaces: facilitation agents which provide the necessary API interfaces for interaction between existing (Web 2.0) and new (Web 3.0) economic models. An example is an AEA with HTTP connection and skill who has the capability to communicate using HTTP.
- Pure software: software agents living in the digital space that interact with interface agents and others.
- Digital data sales agents: software agents that attach to data sources and sell it via the open economic framework. An example can be found here.
- Representative: an agent which represents an individual's activities on the Fetch.ai network. An example can be found here.
Likely short-term applications¶
In the short-term we see AEAs primarily deployed in three areas:
Off-load repetitive tasks: AEAs can automate well-defined processes in different domains such as supply chain, mobility and finance, etc.
Micro transactions: AEAs make it economically viable to execute trades which involve small value transfers. This is particularly relevant in areas where there is a (data) supply side constituted of many small actors and a single demand side.
Wallet agents: AEAs can simplify interactions with blockchains for end users. For instance, they can act as "smart wallets" which optimize blockchain interactions on behalf of the user.
Multi-agent system versus agent-based modelling¶
The multi-agent systems enabled by the AEA framework are technological agent-based solutions to real problems and, although there are some overlap, the framework is not designed from the outset to be used as an agent-based modelling software where the goal is scientific behavioural observation rather than practical economic gain.
Moreover, there is no restriction to multi; single-agent applications are also supported.