Week 31 and 32: Endgame

In these final two weeks the project was tested and some minor adjustments were conducted. The final solution consisted in a system capable of negotiating consumptions based on an optimization algorithm for a house, with capacity of being easily expanded and improved in the future.

The team's final thoughts:

Bruno Ribeiro - It is fascinating to have the opportunity to build the link between the digital world and the physical world. That's what this project allowed us to do by integrating the IoT technology with the versatility of the multiagent systems. It was exciting to be part of this.

Bruno Veiga - This project represents the conclusion of the optimization systems the team had been developing during the year for an energy community and adds an extra focus to the devices inside the house, that ultimately are the ones that the optimization algorithm truly affects and not just the homes in a general way. It was very interesting and challenging and it really helped me understand better multiagent systems and IoT which were fields I did not have a lot of knowledge about.

Carlos Coelho - This last project perfectly completes the previous by allowing the optimization algorithm's results to actually be reflected within a smart home, giving this year a sense of closure. Even though our 2-step negotiation system produced some interesting results, it is stil mostly a proof of concept and needs a lot of work to be market ready. Namely improving the arguments each type of device can use to negotiate as well as tracking the usage patterns of the household and weight them on the negotiation.

Overall, I am very happy with how our 4 projects turned out and the fact that they feel like 4 complementing products that could be used as part of a bigger system, rather than being 4 unrelated academic projects.

Miguel Silva - Our final project of the semester consisted of an energy system that would optimize the distribution of energy between the various house electrical devices. This was achieved through a combination of a multiagent system and some IoT devices. It was fun to see how the house agent was able to manage and distribute energy according to what the agents of the house-devices would argue and/or present as arguments to avoid reducing its energy load. It was particularly interesting as we could actually see a behaviour or an action in our IoT devices however our agent negotiation occurred and would reach an optimal solution.