Notes

We used pure Java to implement the Conference Helper BDI Agent. We did not use any third party libraries. Implementing such an agent presented several challenges to us as we describe below.

First, our agent needs to identify which intentions can it form from the beliefs that it has. Also, these intentions must be related to the desires entered by the user. We accomplish this task by a simple tokenization of given descriptions for the beliefs, intentions, and desires. We split these descriptions into words, and relate the descriptions with a similarity score as shown in the below equation. We picked this equation for the sake of simplicity.

Second, our agent needs to reconsider its plan at each alteration. These alterations can be the user altering his desires or the services/friends notifying the user that some dates changed, events canceled, and so on. Conference Helper BDI Agent reconstructs all of its beliefs, intentions, and future desires in case of an alteration.

Third, for the sake of simplicity, we define friends of the user as service agents, since they are equivalent in implementation level.

Finally, we implement our scenarios as a main Java class. In this class, we first declare initial desires, service agents, and events. Then, we declare our BDI Agent. Last, we simulate each epoch in a for loop, making alterations and printing summaries at each iteration. We pause at each iteration and wait a keyboard input.