Week 3: Cornerstone

Another week has passed and the team is starting to capture the whole picture of the project. After getting to know each other and setting the theme, it was time for the team to gather some knowledge with the expert about the project domain and to design the system architecture.

We scheduled a meeting with our Expert, Professor Zita Vale, in which the objective was to gather all the knowledge so we could begin the conceptualization phase. The idea is to develop an expert system integrated into an aggregator that can manage the energy for some communities giving input on what to do with the energy (sell, energy, minimize consumption, etc). The expert system will know the energy consumption and production of each member of the grid and will evaluate the markets and the flexibility to accomplish two fundamental objectives: decrease energy cost and make better use of the energy produced in excess.

After the meeting, we decided to discuss the ideas and begin to make the initial architecture diagram of our expert system (which you can see in image 1). The architecture is based on the knowledge and the reasoning that we want the system to have.

Image 1: Flowchart of the expert system knowledge base

In the course unit PPROGIA we began a new module about Constraint Logic Programming which is based on the idea that given the constraints about a problem the computer can solve it automatically without being guided directly by the programmer's thought as in a normal program.

In the other unit, ENGCIA, we learned about how to implement uncertainty in an expert system. There are a lot of algorithms but we focused on two: the Theory of Certainty Factors and the Bayes Theorem. Very briefly, the Theory of Certainty Factores is used when the probability of some hypothesis is not supported by statistics (energy consumption of a family), and the Bayes Theorem is used when the hypothesis' is supported by statistics (e.g. Weather Forecast).