Objective: Develop network models for the air transportation systems and a means to test their interaction under conditions of disruption. We can think of the Air traffic system as a series of...

I have uploaded


Objective: Develop network models for the air transportation systems and a means to test their interaction under conditions of disruption. We can think of the Air traffic system as a series of components where each component can fail but there is imperfect information on whether a component has failed. These types of faults are extremely disruptive.   We can map this type of problem to the Air traffic space by comparing it to a decision to A) delay or not delay a flight or B) give away a passenger’s seat to another individual given an inconsistency. This inconsistency would be that the inbound flight (which includes air-crew, passengers etc.) can appear as on time or late depending on where you are in the network. That is, ATC may see that the flight is on time but the gate agent may not have that information.   We don’t care what the consensus is but we know that one all the actors need to agree on a common decision.   There will be a certain amount of delay when communicating the messages between agents. The gate agent and ATC will go back and forth until a consensus is reached. We are not going to try and solve a consensus problem because that is a bit too much. Rather, we are going to say that if no one can agree the flight is delayed/passenger is bumped by default We can examine this from the following perspectives: 1. How much delay between messages can be tolerated before a failure occurs (does ATC need to tell the gate agents within 3 minutes of learning the plane will be delayed?) a. Do messages need to be broadcast to all agents in parallel or can the message be sent in sequence? b. Does the system need to “time out”  mechanic to prevent a failure (If no decision is reached after 20 minutes, a default option is selected) c. What happens if one actor is a bottle neck in communications? IE: no other agents can make a decision until ATC sends a message 2. Is failing to reach a consensus result in worse system performance (ie more delayed flights, passengers have to make more transfers etc ) than choosing the default option. 2. Basically, is indecision worse for the performance of a system than going with less desirable option 2. How many consensus failures can a scheduler tolerate before system performance suffers 1. Does the number of faulty nodes the system can tolerate equal n+1/3 where n is the number of nodes? 3. What happens to the performance of a scheduler when the number of failed nodes is greater than n+1/3 at any one instant?     We are going to assume we are working with airport has 5 networks (ATC, Gate Agents, Fuel/Airplane Servicers, Baggage, Flight Crews). We are going to look at the networks for a single airport. For output I need all commented code with test baseline (no disruption case), disruption test cases and written answers for each questions.