Apparatus and method for resource negotiations among autonomous agents

A method for allocation of resources, said method comprising the steps of:

• employing a communications network for passing messages between a first agent computer and a second agent computer;
• executing programs by a central authority computer, the first agent computer, and the second agent computer;
• receiving a first graph by the first agent computer, the first graph representing a first set of tasks and resources utilized by the first set of tasks;
• receiving a second graph by the second agent computer, the second graph representing a second set of tasks and resources utilized by the second set of tasks;
• utilizing the first graph for determining resources required by the first agent computer to perform the first set of tasks;
• utilizing the second graph for determining resources required by the second agent computer to perform the second set of tasks;
• negotiating between the first agent computer and the second agent computer for resources requires to perform the first and second sets of tasks;
• and
• calculating deadlock avoidance structures based on an internal representation of initialization data.

A computer program product stored in a computer readable medium, said product being operative in a distributed data processing system employing negotiation for allocation of resources, the system including a communications network, said product including:

• a first autonomous agent node receiving a first graph representing a first set of tasks and resources for performing the first set of tasks, the first agent node utilizing the first graph to determine the resources required for performing the first set of tasks;
• a second autonomous agent node receiving a second graph representing a second set of tasks and resources for performing the second set of tasks, the second agent node utilizing the second graph to determine the resources required for performing the second set of tasks;
• the first agent node negotiating with the second agent node for resources required to perform the first set of tasks,
• the first agent node calculating deadlock avoidance structures relating to an internal representation of initialization data and the first graph.