| This dissertation describes a distributed artificial intelligence system developed to solve organic spectroscopy problems. The distributed artificial intelligence system is implemented as distributed agents using a blackboard agent as a mechanism for communication between the agents. The blackboard agent assigns tasks to individual agents, dynamically monitors and balances the workload of each agent, accumulates and assimilates information from the agent nodes and prioritizes the results.; This program has developed over several years and as a result, different phases of development occurred on different hardware and software. The program is written in C and OPS5 with UNIX as the operating system. The initial work was done on an NCUBE hypercube. It was transferred to an Intel Paragon computer. The latest version is running on a series of Dec workstations using PVM (Parallel Virtual Machine) as the interface between the computers.; The input to the system is spectrophotometric data (IR, HNMR, CNMR, and MS) and the output is a list of the most likely structures consistent with the data. |
