Yoshiteru Ishida

Fully distributed diagnosis by PDP learning algorithm: towards immune network PDP model

Based on the strong analogy between neural networks and distributed diagnosis models, diagnostic algorithms are presented which are similar to the learning algorithm used in neural networks. Diagnostic implications of convergence theorems proved by the Lyapunov function are also discussed. Regarding diagnosis process as a recalling process in the associative memory, a diagnostic method of associative diagnosis is also presented. A good guess of diagnosis is given as a key to recalling the correct diagnosis. The authors regard the distributed diagnosis as an immune network model, a novel PDP (parallel distributed processing) model. This models the recognition capability emergent from cooperative recognition of interconnected units

A topological approach to failure diagnosis of large-scale systems

A failure diagnosis model which consists of a set of units that are faulty, a set of measurements to detect a faulty unit, and an incidence matrix representing the binary relation between these two sets is presented. A binary relation is defined by whether or not a fault of one unit is detected by a measurement. Considering this failure diagnosis model as a simplicial complex in topological geometry, the relationship between the diagnostic aspects of the model and the topological properties of the corresponding simplicial complex is studied. The diagnosability of a failure diagnosis model with multiple faults is expressed as a covering property of each simplex in a simplicial complex. The capability of distinguishing a faulty unit by a given number of measurement is determined by examining the global connective structure of the simplicial complex. This analysis of connectivity is done by performing q-analysis on a simplicial complex. It is noted hat all possible combinations of fault patterns are considered under a permissible number of faults. However, in actual systems, only restricted fault patterns are possible in which faulty units are functionally connected with each other, even in multiple fault situations.

An immune network model and its applications to process diagnosis

Immune systems, like neural systems, have a highly sophisticated capability of pattern recognition. However, their recognition and learning mechanisms are quite different from those of a neural system. Neural network models (connectionist models) are information models, which derive partially from the study of the mechanism. In the same manner, an information model is proposed which derives partially from the recognition mechanism of immune systems, and the learning algorithms on the model are studied. This paper also proposes some extensions of the model so that it can be applied to process diagnostic problems. The instrumentation system with this immune network (sensor network) can eliminate abnormal information from faulty sensors autonomously. Other than diagnostic problems, the immune network model is potentially applicable to mutual/group test of a set of VLSI of the same type, computer network immune to the virus.

The immune system as a prototype of autonomous decentralized systems: an overview

We discuss the features of the immune system; its system aspect (compatible with Jernes network view), its process aspect, and its design aspect (compatible with Metchinikoffs self-defining view). Our applications of these three aspects are also presented briefly. Since these features of the immune system agree with the concept of autonomous decentralized systems, we suggest that the immune system can be a typical model for autonomous decentralized systems. In the self-defining process, agents can refer to the self-information, and can interact with the environment. We propose that this self-defining process extracted from the immune system can be a candidate for design paradigm for autonomous decentralized systems where full specification of the total system is not only unavailable but also inadequate. We also suggest that the Internet would be a typical example designed by this paradigm.

A critical phenomenon in a self-repair network by mutual copying

This paper reports a critical phenomenon in a self-repair network by mutual copying. Extensive studies have been done on critical phenomena in many fields such as in epidemic theory and in percolation theory with an effort of identification of critical points. However, from the viewpoints of cleaning up a network by mutual copying, critical phenomena have not much studied. A critical phenomenon has been observed in a self-repair network. Self-repairing by mutual copying is “the double-edged sword” that could cause outbreaks with inappropriate parameters, and careful investigations are needed.

Active noise control by an immune algorithm: adaptation in immune system as an evolution

A new information processing architecture is extracted from immune systems. By focusing on informational features of the immune system (i.e. specificity, diversity, tolerance, and memory), an immune algorithm is proposed. The algorithm proceeds in three steps: diversity generation, establishment of self-tolerance, and memorizing non-self. An agent-based architecture based on the local memory hypothesis and a network-based architecture based on the network hypothesis are discussed. An agent-based architecture is elaborated with an application to control systems where the knowledge about disturbances is not available. An adaptive disturbance neutralizer is formalized and simulated for a simple plant.

An immune algorithm for multiagent: application to adaptive noise neutralization

A new information processing architecture is extracted from the immune system. By focusing on informational features of the immune system (i.e. specificity, diversity, tolerance, and memory), an immune algorithm is proposed. The algorithm proceeds in three steps: diversity generation, establishment of self-tolerance, and memorizing non-self. The algorithm may be used typically to model the system by distributed agents where the system (the self) as well as the environment (the non-self) are unknown or cannot be modeled. Agent-based architecture based on the local memory hypothesis and network-based architecture based on the network hypothesis are discussed. Agent-based architecture is elaborated with the application to an adaptive system where the knowledge about environment is not available. Adaptive noise neutralization is formalized and simulated for a simple plant.

Immune networks for cement plants

Taking inspiration from the immune system mechanism, an application which identifies faulty sensors in the firing section of a cement plant by mutual recognition is presented. Some improvements of existing algorithms are described. Results using simulated data are reported and discussed.<<ETX>>

Spatial strategies in a generalized spatial prisoner’s dilemma

Many strategies, such as tit-for-tat, have been proposed in the iterated prisoner’s dilemma (IPD) in which the prisoner’s dilemma (PD) is carried out repeatedly with two players. A spatial version of the iterated prisoner’s dilemma (SPD) has been studied, where a player at each site plays the IPD game with all the players in the neighborhood. However, the strategies studied in the SPD consider the past actions of a single opponent only. We studied spatial strategies that depend on the configuration of actions taken by all neighbors (as opposed to conventional temporal strategies). Since generosity can be considered as a spatial strategy, we first investigate the generosity required when an action error is involved. We also propose several spatial strategies that outperform many others.

Diagnosability and Distinguishability Analysis and Its Applications

As systems become more complex, it becomes necessary to understand, simplify, and apply fault diagnosis and fault-tolerant design. Although some graph-theoretical diagnostic models such as self-diagnosis model have been studied, the model can not be applied to most systems due to the assumption that each unit has its own testing capability. This paper presents a graph-theoretical diagnosis model expressed by a set of fallible units, a set of measurements, and an incident matrix indicating binary relation between these two sets. Since this model explicitly separates tested units (fallible units) and testing units (measurements), we can discuss diagnostic aspects from both sides. Diagnosability and distinguishability of the model with multiple faults are discussed from combinatorial point of view. Measures of t-fault diagnosability and t-out-of-s diagnosability which was introduced on the self-diagnosis model are discussed. Conditions for these diagnosabilities are expressed by a topological concept of fault distance. The concept of distinguishability is generalized to multiple fault situations called t-fault distinguishability. A lower bound for the distinguishability is obtained by using fault distance. The new concept of s-distinguishability class (s-dc) is presented. This analysis is recommended in the design of systems to attain a required level of diagnosability and distinguishability as well as in the analysis of present systems to investigate their diagnostic aspects. Two application examples are presented: Diagnosability and distinguishability analysis of error-correcting codes, and design of instrumentation systems of large plants with a required level of diagnosability.