Complexity Dimensions Flashcards
Modularity
The extent to which a system can be decomposed into interacting modules that can be understood separately.
What Modularity is used in the project and why?
Hierarchical or modular: The system is modular,and the modules themselves are decomposed into interacting modules.
The project is designed to have several robot machines working in tandem with each other. The robot systems and internal modules will have to work together to run the cooperative environment.
Planning Horizon
The planning horizon is how far ahead the agent considers the consequences of its actions.
What Planning horizon is used in the project and why?
An indefinite horizon planner: An agent that looks ahead some finite, but not predetermined, number of steps ahead.
The robot systems should be able to follow each other based on an unknown defined endpoint.
Computational limits
Sometimes an agent can decide on its best action quickly enough for it to act. Often there are computational resource limits that prevent an agent from carrying out the best action.
What Computational limits is used in the project?
Bounded rationality: where an agent decides on the best action that it can find given its computational limitations.
The actions decided by the agent in the project are limited by the software & hardware limitations. The robot decides on its best course of action based on the numerical accuracy of the data it receives and the hardware processing ability.
What Representation Scheme is used in the project and why?
Features or propositions
The robot agents in the project reason to the features built up from a series of states.
Representation Scheme
The representation scheme dimension concerns how the world is described.
The different ways the world could be to affect what an agent should do are called states.
Preference
Agents act to have better outcomes for themselves. The only reason to choose one action over another is because the preferred action will lead to more desirable outcomes.
What Preference is used in the project and why?
Goals: Either achievement goals to be achieved in some final state or maintenance goals that must be maintained in all visited states.
The H2trac robots have a set final goal state to be reached (the tractor and harvesters follow a path around a field)
Number of Agents dimension
Taking the point of view of a single agent, the number of agents dimension considers whether the agent does.
What Number of Agents dimension is used in the project and why?
Multiple agent reasoning: Where the agent takes the reasoning of other agents into account.
The entire cooperative system requires the robots to be able to work in tandem with each other and take into account the state of the other machines.
What learning dimension is used in this project and why?
Knowledge is learned(from data or past experience): An agent should use data from its past experiences and other sources to help it decide what to do.
The knowledge the agent uses is learned from observed data from its environment and the real time input data provided.
Learning
Learning typically means finding the best model that fits the data. Sometimes this is as simple as tuning a fixed set of parameters, but it can also mean choosing the best representation out of a class of representations.
What Sensing uncertainty dimension is used in this project and why?
Fully observable is when the agent knows the state of the world from the observations.
The H2trac robots need to be able to assert the state of the world from its observations to be able to run its cooperative system.
