Trying to implement singleserver with RPDEVS

• Changes in RPDEVS:
  • only one transition function δ, called at external and internal events
  • output function λ depends on state and input, called before any transition
  • different behaviour of coupled system to cope with Mealy situations
    • first: cascade of λ calls throughout all components
    • leads to changes of inputs of the components and corresponding outputs
    • finally: states change with δ
  • algebraic loop → λ cascade doesn't terminate
• Singleserver in RPDEVS:
  • implements original state diagrams without transitional states
  • λ and δ functions much simpler than in PDEVS
  • testing with PowerRPDEVS → model aborts at t = 1 (too many λ steps)
  • problem of λ iteration
    • cascades of outputs are computed
    • states of the components don't change
    • e.g. queue gets blocking input, but doesn't change state
    • behaviour is not consistent with idea of state diagram
  • RPDEVS is a great idea, but the λ iteration has the wrong semantics