This paper presents a fast convergence block adaptive filter in which the filter weights are adjusted based on the conjugate gradient method. The proposed algorithm permits the use of the fast convolution in accordance with the overlapsave method using the FFT, so that it can reduce the computational complexity to O(N log N + N) for N taps FIR filters. Some computer simulations show the faster convergence property of the proposed method than the conventional algorithms such as well-known LMS type techniques.

This paper treats mutual exclusion of a single shared-resource in distributed autonomous environments. The most important property of the autonomous network treated in this paper is its membership variability, that is, frequent occurrence of entries of new nodes and exits of old nodes. Thus, when the network is large-scale, it is not possible for each node to keet up the information of all other nodes. We in this paper design a mutual exclusion algorithm for distributed environments of autonomous nodes based on Chandy-Misra protocol for Dining Philosopher (diners) problems, which realizes a distributed implementation of the token ring method. We consider requirements of the communication topology that makes mutual exclusion possible, and propose entry and exit protocols for each node to perform them individualistically and autonomously.

A new concept of fuzzy tensions and fuzzy potentials is proposed as a mathematical tool for qualitative analysis of network problems. In this paper we first describe modeling of curriculum networks and fail-safe networks in which the concept of fuzzy tensions appears naturally as a well-balanced curriculum plan or reliability assignment. Algebraic properties of fuzzy tensions and potentials are thoroughly investigated and consistency of specifications is shown to be expressible by a simple condition on circuits of the underlying network.

The important problems in IIR adaptive equalizers are stability of the adaptive IIR filters and unimodality in the error surface. In this report we propose a new IIR adaptive equalizer which assures the unimodality and the stability.

We discuss properties of acyclic graph evolution driven by node-firing. The research background and basic concepts of acyclic graph evolution are from the mutual exclusion problem in distributed environments. We proposed in our previous work a mutual exclusion protocol which is based on the notion of evolution trajectories of acyclic graphs. In this paper, we analyze firing concurrency and periodicity of the acyclic graph evolution, from graph theoretical point of views, and investigate topological conditions for assuring the number of firable nodes below a some fixed constant, at any instance of the evolution trajectory. A marked graph, a subclass of Petri nets, is often utilized as a proof tool in analysis.

In this paper we consider a sex-fair matching in the stable marriage problem. The sex-fair stable matching defined in this paper has a property that the sum of partner's ranks of individual men in their preference lists is as close as possible to the sum of partner's ranks of individual women in their preference lists. However, the sex-fair stable marriage problem is known as one of important open problems in the stable marriage problems. The main result of this paper is to propose a method for adapting a genetic algorithm (GA) to the sex-fair problem to find the sex-fair matching effectively. In the method we transform the sex-fair marriage problem into a graph problem whose decision version is shown to be NP-complete. The graph problem representation is suitable for GA application, that is, easier coding, easier and more effective definition of genetic operators. Computer experiment reports effectiveness of the GA solution.