Hitachi, Ltd.

Seiji Sakaguchi

Motohisa Funabashi

Shoichi Masui

Shunichi Tano

Hiroshi Tomita

Hiroshi Koike

Noriko Sasaki

Shiochi Masui

Actual data which is satisfied by a condition string which can be judged within actual data is stored in each .alpha. memory, and combined actual data is generated from each of the .alpha. memories by a combined actual data generator, and the inter-actual data condition of the generated combined actual data is evaluated by an inter-data condition evaluation for actual data so as to specify an instance with no conflict resolution process being performed and to reduce the memory capacity necessary for the evaluation process by a production system.

Method and system for a production system

In an inference method using a rule, a rule condition part network and rule consequence part network are generated, an integral network of both the networks is used for inference. With this information method, the inference is performed at high speed by setting a shortcut arc so as to process the inference along the pattern matchable network portion. Further, use of the integral network of the rule condition and consequence parts enables not only forward and backward inferences but also a bidirectional inference with intimate couple between forward and backward inferences.

Fast method for a bidirectional inference

A method for generating a network to cope with a conditional part or an IF part having a comparison between variable data and a logical sum between comparisons which appear in a Rete network adopted in a production system using the knowledge technology scheme. In association with an intra-node, a candidate node is disposed in another conditional clause referencing a variable, and a data which was previously accepted but which did not satisfy the conditions or a data most likely to pass through the network is stored in the candidate node in advance. For the intra-node in the object processing, the value of the candidate node is loaded for judging the conditions.

Method for judging whether conditions are satisfied by using a network

A knowledge-based system comprising a plurality of knowledge processors each of which includes a knowledge base and performs a reasoning process for replying to a request from a user, and a file server which is connected to the knowledge processors through a transmission line. When each knowledge processor cannot resolve the user's request with the knowledge base of its own, it sends a request message to the other processor or processors through the transmission line. Reply messages from the other knowledge processors complying with the request message are managed in correspondence with the requests by the file server which receives the reply messages, and when a predetermined condition has been established, a reply is given to the knowledge processor which is the source of the request. Thus, each knowledge processor replies to the user's requests while automatically switching the reasoning process in a local mode which utilizes only the knowledge base of the processor itself and the reasoning process in a multiple or global mode which is based on cooperation of the other knowledge processor or processors.

Knowledge-based system having plural processors

A dynamic network reconstruction method in which the node processing information such as a node processing volume and a ratio at which a node is satisfied is collected for all nodes to maintain an efficient network even when the characteristic of an object described in a condition clause is changed in a Rete network of a production system using the knowledge technology and the network is changed based on the node processing information.

Growing community of inventors

Yokohama, Japan

Seiji Sakaguchi