Coursework requirements
The second coursework builds on Coursework 1, and uses the same 2D environment and task. However, in the second coursework, multiple cooperating agents must collect waste from stations and dispose of it in wells. The goal of the agents is to dispose of 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
· tasks persist until they are achieved (a station has at most one task at any time)
· the maximum amount of waste that must be disposed of in a single task is 1,000 litres
· wells can accept an infinite amount of waste
· refuelling points contain an infinite amount of fuel
· in each run, there is always a refuelling station in the centre of the environment
· a run lasts 10,000 timesteps
· 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
· the agent can take waste from a station and dispose of it in a well
· moving around the environment requires fuel, which the agent must replenish at a fuel station
· the agent can carry a maximum of 100 litres of fuel and 1,000 litres of waste
· the agent starts out in the centre of the environment (at the fuel station) with 100 litres of fuel and no waste
· the agent moves at 1 cell / timestep and consumes 1 litre of fuel / cell
· filling the fuel and waste tanks and delivering waste to a well takes one timestep and no fuel
· if the agent runs out of fuel, it can do nothing for the rest of the run
· the success (score) of an agent in the task environment is determined by the amount of waste delivered.
The task environment should not be modified or extended. You must implement a multi-agent system that completes the task in the specified task environment, and must include in your final report an evaluation of theperformance of your agents (average score over at least ten runs divided by the number of agents used).
The objective
Objective is to investigate agent architectures and coordination mechanisms for water collection and delivery strategies, to decide
– which agent architecture(s) to use given the task environment
– specialised agent vs homogeneous
– deciding how to allocate tasks to agents
Main problem
• adding multiple agents introduces a new problem of coordination – how should agents explore the environment, and what information
should they share with other agents
• how do agents decide which tasks to perform and how to (who should) perform them so that
– two agents don’t try to perform the same task, and the largest number of tasks are achieved
/docProps/thumbnail.jpeg