Although production systems are widely used in artificial intelligence (AI) applications, they are seen to have certain disadvantages in terms of their need for special purpose assistance software to build and execute their knowledge-bases (KB), and in the fact that they will not run on any operating system (platform dependency). Furthermore, for AI applications such as learning assistance systems, there is a strong requirement for a self-adaptive function enabling a flexible change in the service contents provided, according to the user. Against such a background, a Java based production system (JPS) featuring no requirement for special purpose assistance software and no platform dependency, is proposed. Furthermore, a new self-adaptive Java production system (A-JPS) is proposed to realize the "user adaptation" requirement mentioned above. Its key characteristic is the combination of JPS with a Causal-network (CN) for obtaining a "user profile". In addition, the execution time of the JPS was studied using several benchmark problems with the aim of comparing the effectiveness of different matching algorithms in their recognize-act cycles as well as comparing their performance to that of traditional procedural programs for different problem types. Moreover, the effectiveness of the user adaptation function of the A-JPS was studied for the case of a CN with a general DAG structure, using the experimental KB of a learning assistance system.

This paper presents a new hierarchical scheduling method for a large-scale manufacturing system based on the hybrid Petri-net model, which consists of CPN (Continuous Petri Net) and TPN (Timed Petri Net). The study focuses on an automobile production system, a typical large-scale manufacturing system. At a high level, CPN is used to represent continuous flow in the production process of an entire system, and LP (Linear Programming) is applied to find the optimal flow. At a low level, TPN is used to represent the manufacturing environment of each sub-production line in a decentralized manner, and the MCT algorithm is applied to find feasible semi-optimal process sequences for each sub-production line. Our proposed scheduling method can schedule macroscopically the flow of an entire system while considering microscopically any physical constraints that arise on an actual shop floor.

In loudspeaker-based 3D audio systems, there are some acoustic crosstalk cancellation methods to enlarge the 'sweet spot' around a fixed listener position. However, these methods have common defect that most of them can be applied only to the specific narrow frequency band. In this letter, we propose the more robust acoustic crosstalk cancellation method so that we can cancel the crosstalk signal in far wider frequency band and enlarge 'sweet spot. ' For this goal, we apply a sum and difference filter to the conventional three loudspeaker-based 3D audio system.

We describe a method of compensating temperature fluctuation by a linear-time-warping processing in a sound reproduction system. This technique is applied to impulse responses of room transfer functions, to achieve a high-quality sound reproduction system, particularly one that treats high-frequency components. First, the impulse responses are measured before and after temperature fluctuation, and the former are converted to the latter by the proposed process. Next, we design inverse filters for the system, and evaluate the improvement of the reproduction accuracy and spectrum distortion. By the compensation method, we can improve the reproduction accuracy at any frequency. Moreover, we propose an adaptive algorithm for the estimation of a suitable warping ratio, using the observed signal of reproduced sound obtained at only one control point. Using the proposed algorithm, we can improve the reproduction accuracy at each control point by about 14 dB, in which a difference in temperature is 1.4.

A new architecture and methods for an enhanced autonomous decentralized production system (EADPS) are described. This EADPS was developed to ensure high flexibility of production systems consisting of intelligent devices based on the autonomous decentralized system model and to guarantee the time used for communication to simultaneously maintain high productivity. The system architecture of the EADPS guarantees the time by managing groups of nodes and the priorities in these groups. A bit-arbitration method is used to prevent collision of messages. The nodes autonomously check the waveforms in the network and terminate transmission when the nodes with a higher priority are transmitting. A parallel-filtering method is used to speed up message acceptance. The nodes check the identifiers of the messages using parallel-filtering circuits and each node determines autonomously where a message should be accepted or not. Implementing the system architecture and these methods as circuits and integrating the circuits into a chip using system LSI technologies resulted in low-cost implementation of the system. Experimental evaluation demonstrated the effectiveness of this system.

We discuss human collaborative discovery processes using a production system model as a cognitive simulator. We have developed an interactive production system architecture to construct the simulator. Two production systems interactively find targets in which the only experimental results are shared; each does not know the hypothesis the other system has. Through this kind of interaction, we verify whether or not the performance of two systems interactively finding targets exceeds that of two systems independently finding targets. If we confirm the superiority of collaborative discovery, we approve of emergence by the interaction. The results are: (1) generally speaking collaboration does not produces the emergence defined above, and (2) as the different degree of hypothesis testing strategies that the two system use gets larger, the benefits of interaction gradually increases.

A new method of sequential control has been developed in order to increase the productivity and flexibility of production systems. This advanced sequential control (ASC) method is proposed for sequential control systems based on the autonomous decentralized system (ADS) architecture. The ADS defines the system software and message formats and makes it easy to expand the number of devices and software modules. The ASC method increases productivity because it minimizes the processing and adjustment times of production lines by adjusting the starting times of production processes automatically. Experimental evaluation results of the ASC method showed that it increases the productivity of production systems. It is also applied to an actual production system and the results are reported.

Because the match phase in OPS5-type production systems requires most of the system's execution time and memory accesses, we proposed hash-based parallel production systems, CPPS (Clustered Parallel Production Systems), based on the RETE algorithm for distributed memory parallel computers, or multicomputers to reduce such a bottleneck. CPPS was effective in speeding up the match phase, but still left room for optimizations. In this paper, we introduce software cache techniques to memory nodes in the CPPS as one of the optimizations, and implement it on a multicomputer, nCUBE2. The benchmark results show that the CPPS with the software cache is about 2-fold faster than the original, and more than 7-fold faster than the simple hash method proposed by Acharya et al. for a large scale problem. The speed-up can be attributed to decreased communication costs.

This paper describes, from a system architectural viewpoint, how knowledge-based technologies have been utilized in developing EXLOG (an LSI circuit synthesis system) and SOFTEX (a software synthesis system) inside the authors' projects. Although the system architectures for EXLOG and SOFTEX started from the same production systems, consisting of transformation rules in the middle of the 1980's, both branched off in different directions in the 1990's. Based on experiences with EXLOG and SOFTEX, the differences between LSI and software design models are discussed, and the future directions are indicated for the knowledge-based design system architectures.

In this paper we propose an architecture and an algorithm for the parallel execution of OPS5 production systems. It is known that current OPS5 production system interpreters spend almost 90% of their execution time in the match step. Thus, in this paper we focus on the speedup of the match step. The match algorithm used in OPS5 is called Rete and the algorithm uses a special kind of a date-flow network compiled from the left hand sides of rules. To achieve the maximum degree of parallelism of a given OPS5 program by as few processors as possible, the proposed parallel machine uses loosely coupled multiprocessors. Parallel machines designed for fine-grain parallelism, such as DADO, also use loosely coupled multiprocessors. However, the proposed machine differs from such machines at the following points: use of powerful processors which have large amounts of memories and small cycle times; use of a shared Rete network (parallel machines designed for fine-grain parallelism use an unshared Rete network); high hardware utilization. Basic ideas of the proposed parallel machine are as follows. (1) Use of a modified Rete network in which node sharing is used only for constant-test nodes and each memory node is lumped with the child two-input node. (2) Static allocation of the nodes of the modified Rete network onto processors. (3) Partition of the set of processors into three subsets: constant-test node processors, two-input node processors and conflict-set processors. (4) Use of a ring network for the interconnection network among two-input node processors. In addition to an architecture for parallel execution of OPS5 production systems, we propose a scheme for mapping the modified Rete network into the proposed architecture. The results of simulation experiments showed that the proposed architecture is promising for parallel execution of OPS5 production systems.

A job shop system typically seen in flexible manufacturing systems (FMS) is a system composed of a set of machines and a various kind of jobs processed with the machines. A production system of semiconductor fabrication is an example of job shop systems, which has main features of repetitive processes of one part and set-up times required for machines processing different types of parts. On the other hand, timed Petri nets are used for modelling and analyzing a wide variety of discrete event systems. There are many applications of timed Petri nets to the scheduling problems of job shop systems. The performance evaluation and steady state behaviors are studied by using the maximum cycle time of timed marked graphs. The aim of this paper is to propose a new model for production systems including repetitive processes and set-up time requirements which enables the quantitative analysis of real time system performance. In job shop systems such as a semiconductor fabrication system, it takes considerable amount of set-up time to prepare different types of chemical reactions and the model should take account of a set-up time for each machine. We focus upon the relationship between facility utilization factor and production cycle time in the steady state. In the proposed model, the minimum total set-up time can be attained. Quantitative relationship between utilization factor and production cycle time is derived by using the proposed model. A utilization factor of a system satisfying a given limit of the cycle time is evaluated, and the improvement of the utilization factor is considered. Conversely, we consider the improvement of the cycle time of a system satisfying a given limit of utilization factor.

This paper discusses the design, implementation, and performance of a bus-connected multiprocessor, called Presto, for a Rete-based production system. To perform a match, which is a major phase of a production system, a Presto match scheme exploits the subnetworks that are separated by the top two-input nodes and the token flow control at these nodes. Since parallelism of a production system can only increase speed 10-fold, the aim is to do so efficiently on a low-cost, compact bus-connected multi-processor system without shared memory or cache memory. The Presto hardware consists of up to 10 processisng elements (PEs), each comprising a commercial microprocessor, 4 Mbytes of local memory, and two kinds of newly developed ASIC chips for memory control and bus control. Hierarchical system software is provided for developing interpreter programs. Measurement with 10 PEs shows that sample programs run 5-7 times faster.