Example 3: Queueing system
    
      - Basic system:
        
          - queueing system with generator, queue and server
            
              
 
              - blocking signal from server to queue
 
              - → queue changes state between "Free" and
                "Blocked"
 
            
           
          - constant values tG = 1, tS =
            1.5
 
        
       
      - Model in PowerRPDEVS:
        
          - components described as atomics in RPDEVS
            
              - RPDEVS = revised version of standard PDEVS
 
              - allows for Mealy components
 
              - simulator has to change, includes "λ iterations"
 
            
           
          - PowerRPDEVS simulator
            
              - components described in C++
 
              - construction of the complete example in a
                graphical environment
 
            
           
          - run →
            
              - Error at t = 1.
                maximum number of λ steps has been reached.
                illegitimate model due to a non-resolvable algebraic loop. 
            
           
          - solver follows outputs through the complete model
            
              - iterates, until a steady state is reached
 
            
           
          - does not happen here:
            
              - queue sends entity to server
 
              - server sends "blocked" to queue
 
              - queue changes state to "Blocked"
 
              - queue doesn't send entity
 
              - server is not blocked
 
              - queue sends entity to server ...
 
            
           
        
       
      - Mathematical problem:
        
          - states and output values undefined at t = 1
 
          - more concrete:
            
              - several states and outputs at the same time
 
            
           
        
       
      - Reality check:
        
          - signal transports and state changes need a small time
            delay
 
          - steps in event cascades happen at increasing times
 
          - problem for modeling
            
              - lots of unknown tiny parameters
 
            
           
        
       
      - Extended Model:
        
          -  basic idea
            
              - use hyperreal numbers *R for time values
 
              - add infinitesimal delays for signal transport and
                state changes
 
            
           
          - often a default delay ε is sufficient almost
            everywhere
 
          - simulator has to be adapted accordingly
 
          - queueing example is easily adapted
            
              - delay of state change in queue has to be enlarged
                to guarantee correct order of events