Each student will write a 2-3 page proposal. The project could be to extend the language A for a particular application (e.g. web services, workflows, BIoworkflows, cellular interaction networks), or to extend the language for goals or ditrectives, or to have new kind of agent architecture, or to develop algorithms for particular tasks of agents, or to prove a result, or to develop an agent in a particular domain.
Your proposal should have enough details about what you plan to do. For example, if you will be developing an agent then you need to describe in English (i) What the agents goal(s) are. (i.e., what is it that the agent is supposed to do or achieve.) (ii) What are some of the individual actions than the agent can do; what is the environment; and what are the impacts of the agent's actions on the environment. (iii) Other issues, such as how autonomous the agent should be in terms of how much planning it should do and how reactive it should be. Following are some web links that may be of help.
If your proposal is not approved by me you must submit another proposal at the earliest. The sooner you have your proposal approved, the more time you will have to work on it. Feel free to email me your project proposals. I will prefer simple ascii rather than attachments.
Note that I added two new web sites about policies. Policies are basically a form of reactive control. These two sites may give you some more ideas. Also the book by Subrahmanian et al. mentioned in the reading list has several agent examples.
In this phase, you will report your intermediate progress. For example, if you are developing an agent then you need to do the following:
Precisely describe how the environment is represented. State the relationships between the fluents in the environment using static causal laws and/or `always' statements.
Precisely describe what the agent's actions and and what the environment's actions are. When are the agent's actions executable. Which environmental action can happen when.
What is the impact of the actions on the environment?
State the goal formally using temporal logic; the notion of maintainability; or in formulation of your own.
Now what is your plan about how the agent will achieve/maintain its goal? Do you want to have reactive control? A deliberative architecture? Expand on your thoughts.
Complete your project. For example, if you are developing an agent then, if the control of your agent is written by you then formally prove that the agent will indeed achieve its goal if it follows your control. If the control uses a planner, give details about your architecture, what planner is used, etc.