Other communication paradigms with input and output
In addition to what is discussed above, there are a few other interesting communication paradigms that distinguish input and output. We mention here the following. For each communication paradigm, there is a corresponding algorithm for reachability analysis.
- Input-output pairing with rendezvous interactions (normally using the "!" - "?" notation, as supported by UPPAAL) - see above.
- The outputting component initiates the interaction (as in the IOA model): directly coupled (atomic) input and output transitions in each of the communicating components - see above.
- Using partially defined models, one can express at the same time the guarantees provided by the component behavior, as well as the assumptions that the component may make about input interactions coming from the environment.
- Moore-type state machines - or - using shared variables for communication: The output of one machine depends on the state of the machine (and therefore is a constant value while the machine remains in its current state) - input transitions have constraints about the input provided by machines in the environment (in the form of the values of their output variables). Therefore the possible (input) transitions depend on the current state of the machines in the environment. There are three modes of communication:
- synchronous communication: There is a global clock, and all components perform a transition when the clock ticks. There would be a deadlock if any of the components cannot perform a transition, given the current values of its inputs from the different machines in its environment.
- asynchronous communication: no further coordination between the transitions of different components, however, interleaving semantics is assumed (only a transition of a single component at a time).
- synchronous communication with stuttering: This is like synchronous communication, although a component is not obliged to perform a transition when the clock tick arrives (one says that the component does a stuttering step). Therefore the next transition of the composed system includes the joint synchronous transition involving all components (as in the case of synchronous communication), the individual transitions of any particular component, as well as sets of transitions belonging to different components (if there are more than two components in the system).
- Give some example ***
- Control communication: a special form of communication by which real-time controllers communicate with the "plant" that they control (similar to the IOA communication model) by people working on the synthesis of controllers that communicate with a plant to ensure that the interactions of the plant satisfy certain requirements (objectives). The interactions between the controller and the plant are classified as:
- hidden, or
- visible. The visible interactions are classified as
- controllable: the controller can prevent that these interactions are performed by the plant. This is equivalent to considering that they are rendezvous interactions. Note that the controller cannot enforce such an interaction when the plant is in a state where it is not willing to do such interaction.
- uncontrollable: this means that the plant may perform these interactions even if the controller would like to prevent these interactions from happening. This appears to be equivalent to considering that these interactions are output interactions initiated by the plan (same semantics as for IOA systems).
- Reference to controller synthesis ***
- Give some example ***
Created: September 22, 2014