See package description for more information.
start();
. Agents may send and receive messages to communicate with other agents and/or hosts. An agent has a reference to its execution environment, a so called container, which it may use to initiate actions, e.g. send a message or require the de.uniol.informatik.vlba.purplebee.transfer to another host. The agent does only interact with the container but not with any other component in the application or the network.
@author dluebke
@version 1.0.3 (2003-11-26)
@see de.tuclausthal.informatik.winf.mobileagents.container.Container
@since 1.0
The agent tracks its intentional state using a plan tree, and maintains dialogue state by keeping track of a focus within that tree. At each turn, it tries to find an elaboration that adds to the plan tree as near the focus as possible, and has a primitive task as a leaf. It then adds this to the tree and executes the task (which becomes the new focus).
This class provides no UI; see {@link AgentConsole} and {@link AgentServer}, which both build (different) UIs on top of it. @author Dan Schulman @version $Id: Agent.java 1009 2011-01-21 21:02:36Z schulman $
Instances of this class are immutable and so can be shared freely.
@author Andy JacksonAgent
class is the common superclass for user defined software agents. It provides methods to perform basic agent tasks, such as: ACLMessage
objects, both unicast and multicast with optional pattern matching. Agent
subclasses, adding specific behaviours as needed and exploiting Agent
class capabilities.
@author Giovanni Rimassa - Universita' di Parma
@author Giovanni Caire - TILAB
@version $Date: 2011-06-06 09:17:37 +0200(lun, 06 giu 2011) $ $Revision: 6414 $
Implementations that provide the set of capabilities must clearly define the names of these capabilities and their values.
Capability names that are implementation private must be defined so using appropriate naming schemes. Trinidad private capability names are prefixed using "-adfinternal-xxx", and such capability names (and their values) may change at anytime (and not guaranteed to be supported in future releases).
This class holds some state that relates to the Agent and in addition some methods may be called on the agent. destroy(), invokeMethod() and refresh() are actually called by a proxy class AgentProxy. AgentProxy is actually an interface that is implemented by Console (as that's where all the JMS stuff is), we use this approach to avoid introducing a circular dependency between Agent and Console.
The Console application maintains a list of all known remote Agents. Each Agent is represented by an instance of the Agent class:
The following diagram illustrates the interactions between the Console, AgentProxy and the client side Agent representation.
@author Fraser Adams
Agent
is a server plugin that can be initialised, started and destroyed along with the MuleContext itself. Agents can initialise or bind to external services such as a Jmx server.
Agent
object represents a tracing agent used to monitor events on a connection. Its primary responsibilities are to create Trace
objects that are attached to a specific socket channel. When any event occurs on that channel the trace is notified and can forward the details on to the agent for analysis. An agent implementation must make sure that it does not affect the performance of the server. If there are delays creating a trace or within the trace itself it will have an impact on performance. @author Niall Gallagher @see org.simpleframework.transport.trace.Trace
The following features are supported:
The following features are supported:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|