CSCE 475/875

Final Project Assignment:

Local Decisions vs. Global Coherence

Assigned:  September 23, 2013 

Due:  December 16, 2013

 (Project 5 minutes late will not be accepted)

 

A Supplement: An Example, Continued

 

This supplement is to give you an example of how changes are to be reported after your proposal has been reviewed.   

 

Comment #1:  In your problem statement, you indicated that each mine is set to go off on its own within a specific time.  However, in your proposal, the range of these values is not stated.

 

Response:  Each mine is set to go off on its own after 1-500 ticks, assuming that each agent can only take at most one step per tick.  The actual value will be randomly generated for each mine using a uniform distribution.

 

Comment #2:  How are you going to simulate each mine?  Perhaps, you will need to create a second type of agents: “mine agents”.  These agents will be just dummy agents.  Each has a location and a number that indicates the number of cycles to explosion.  And the number ticks down until the agent explodes.  And then each agent also keeps track of its status: ticking or exploded.  Something like that.

 

Response:  Yes, we will create a second type of agents to deal to “represent” mines.  These agents will not be able to communicate with other agents at all.  However, a mine sweeper agent will be able to query a mine agent to find out the “time” to explosion.

 

Comment #3:  In your experiments, you did not indicate how you were going to analyze your results to obtain a report on local decisions vs. global coherence.  The final project assignment (under “Experiments and Report”) has very specific items to be included in the summary of your simulation runs.

 

Response:  Making local decisions is reflected in the number of agents that an agent contacts throughout its lifetime in our experiments, and the amount of messages sent by the agent.  The smaller the number of agents that an agent contacts and the lesser the amount of messages sent by the agent indicate that the agent is likely making more local decisions.  Thus, in our experiments, we will also collect data on these two numbers and report on how they relate to the emergent behavior (defined as effectiveness and efficiency as previously stated in our proposal).