G53DIA Coursework 1 Description
The first coursework involves the specification, design and implementation of a simple agent.
Coursework Requirements
The problem consists of a 2D environment, in which a single agent must collect and dispose of waste (CO2) from stations, e.g., carbon capture and storage. Stations periodically generate tasks requests to dispose of a specified amount of waste. The environment also contains a number of wells where waste can be deposited. The goal of the agent is to collect as much waste as possible in a fixed period of time.
Task Environment
The standard task environment is defined as:
the environment is an infinite 2D grid that contains randomly distributed stations, wells and refuelling points
stations periodically generate tasks requests to dispose of a specified amount of waste (max 1,000 litres)
tasks persist until they are achieved (a station has at most one task at any time)
wells can accept an infinite amount of waste
refuelling points contain an infinite amount of fuel
the agent can carry a maximum of 200 litres of fuel and 1,000 litres of waste
the agent can see any stations, wells and refuelling points within 20 cells of its current position
if a station is visible, the agent can see if it has a task, and if so, how much waste is to be disposed of
move actions take one timestep, consume 1 or 2 litres of fuel, and may or may not move the agent 1 cell (depending on the type of move action)
filling the fuel and waste tanks and disposing of waste in a well takes one timestep (and consumes no fuel)
some actions may fail, leaving the agent in the same state as before the action
the agent starts out at a fuel station with 200 litres of fuel and no waste
a run lasts 10,000 timesteps unless the agent runs out of fuel, in which case the
run is terminated
the success (score) of the agent is determined by the total amount of waste
collected
The task environment should not be modified or extended. All other decisions regarding software design and implementation strategy are up to you. However, you will be given guidance and feedback on your project in individual tutorials.
You must implement an agent that completes the task in the specified task environment and include in your final report an evaluation of the performance of your agent (average score over at least ten runs).
Resources
A Java package (package uk.ac.nott.cs.g53dia) is provided as a starting point for your project work. This provides an implementation of the standard task environment (in the library package) and a very basic DemoTanker (in the agent package) that chooses actions at random. A basic GUI is also provided (in the simulator package).
You must not modify the package hierarchy or any of the code in the library package. You may modify the code in the simulator package, but any modifications made to this package will not be considered in marking. All the code you submit must be in the agent package.
The project work is supported by tutorials that cover the design and implementation of your agent and the use of the Java package.
Assessment
Assessment is by means of project work and reports.
What is Required
A report describing your agent and the associated source code. You must include in your report an evaluation of the performance of your agent in the standard task environment (average score over at least ten runs). Your report must not exceed 2,000 words excluding references.
The source code should be submitted as a zip archive of the agent package (Java source files only).
Do not change the package structure.
Do not include files from the library or simulator packages in the zip archive.
Do not submit any code that is not in the agent package.
Do not include any code (e.g., debugging code) that writes to stdout.
If you do any of these things you will lose marks.
Coursework 1 constitutes 50% of the assessment (plus 5% for the environment classification = 55%).
Marking
Marking will be based on:
the capabilities of the implemented agent system, including the quality of the specification, design and implementation;
the degree to which the specification, design and implementation are clearly documented in the report, including any relevant background material used in the design and implementation of the agent; and
clarity of presentation in general (including grammar, spelling and punctuation).
Extra credit will be given for submissions that demonstrate a clear understanding of the relationship between the specified task environment and the architecture of the implemented system.
Assessment criteria and typical requirements for specific grades are available on moodle.
Submission
Submissions are due on Thursday 14th of March at 15:00.
Reviews
There are no reviews yet.