Abstract

The present paper introduces some basic concepts and terminology of distributed planning, and describes a novel coordinator‐coworker structure—a control mechanism that can dynamically reorganize the planning system in response to the tasks at hand. Coordinators get either self‐appointed in unambiguous situations or selected through a unique and inexpensive nomination/confirmation process. The matching of tasks and coworkers is accomplished with minimum message passing and processing, by using appropriate meta‐knowledge about the status of the network of nodes. Connections can be initiated by both coordinators and coworkers, leading to optimum load balancing. A hierarchically organized set of distributed scratch pads supports the coordinator‐coworker mechanism. It is the medium for information exchange and also provides the basis for self‐repairing in a graceful degradation mode whenever individual nodes become inoperational. We discuss the problems of air‐traffic control and describe the kernel design of ...

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call