Week 29: Continuing the implementation

In this week, the team continued the implementation of the multi-agent system and started to create two iot devices that will be used to simulate consumption and generation of the house in study.

Regarding the multi-agent system, it is now capable of simulating 24 periods. The second phase of negotiation referred in the last post has also been implemented in the fridge agent as well as the two agents that will communicate with the iot device. This negotiation will consist in the device agents sending arguments, which consist in characteristics of their devices in order to convince the house agent to let them operate with the consumption level they requested. In this phase, the house agent is also able to ask for certain aspects of the device agents in order to help him decide.

In terms of IOT, the team is using two esp8266 to simulate consumption and generation. The generation device communicates with MQTT and uses beside the esp8266, two buttons, a potentiometer and 4 leds. This device has two ways of simulating generation, a profile based generation of four levels that uses the leds and a button, and a dynamic continuous way that uses the potentiometer to change the generation value. The consumption device consists of several leds that will act as a consumption bar, where if the more leds are on, then there is a bigger consumption value. This iot will use HTTP requests to communicate with the agent and it is adapted to work with a flexible number of leds, meaning that the consumption bar will be able to use the number of leds chosen by the user. Despite the devices implementation being in advanced stages, there is still the need to further test them, as well as integrate them within the multi-agent system.

The classes of this week focused on communication protocols in the Smart Environments class, and properties and architectures in the Multi-agent class.