• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.39 no.4
• /
• pp.81-89
• /
• 2016

Distributed genetic algorithm (DGA), also known as island model or coarse-grained model, is a kind of parallel genetic algorithm, in which a population is partitioned into several sub-populations and each of them evolves with its own genetic operators to maintain diversity of individuals. It is known that DGA is superior to conventional genetic algorithm with a single population in terms of solution quality and computation time. Several researches have been conducted to evaluate effects of parameters on GAs, but there is no research work yet that deals with structure of DGA. In this study, we tried to evaluate performance of various genetic algorithms (GAs) for the famous symmetric traveling salesman problems. The considered GAs include a conventional serial GA (SGA) with IGX (Improved Greedy Crossover) and several DGAs with various combinations of crossover operators such as OX (Order Crossover), DPX (Distance Preserving Crossover), GX (Greedy Crossover), and IGX. Two distinct immigration policies, conventional noncompetitive policy and newly proposed competitive policy are also considered. To compare performance of GAs clearly, a series of analysis of variance (ANOVA) is conducted for several scenarios. The experimental results and ANOVAs show that DGAs outperform SGA in terms of computation time, while the solution quality is statistically the same. The most effective crossover operators are revealed as IGX and DPX, especially IGX is outstanding to improve solution quality regardless of type of GAs. In the perspective of immigration policy, the proposed competitive policy is slightly superior to the conventional policy when the problem size is large.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.3
• /
• pp.261-268
• /
• 2002

This paper represents that auto tunings of membership functions and weights in the fuzzy neural networks are effectively performed by immune algorithm. A number of hybrid methods in fuzzy-neural networks are considered in the context of tuning of learning method, a general view is provided that they are the special cases of either the membership functions or the gain modification in the neural networks by genetic algorithms. On the other hand, since the immune network system possesses a self organizing and distributed memory, it is thus adaptive to its external environment and allows a PDP (parallel distributed processing) network to complete patterns against the environmental situation. Also, it can provide optimal solution. Simulation results reveal that immune algorithms are effective approaches to search for optimal or near optimal fuzzy rules and weights.

• Journal of Electrical Engineering and information Science
• /
• v.3 no.2
• /
• pp.151-157
• /
• 1998

Free-space optical interconnection networks can be classified into two types, space variant and space invariant, according to the degree of space variance. In terms of physical implementations, the degree of space variance can be interpreted as the degree of sharing beam steering optics among the nodes of a given network. This implies that all nodes in a totally space-invariant network can share a single beam steering optics to realize the given network topology, whereas, in a totally space variant network, each node requires a distinct beam steering optics. However, space invariant networks require mechanisms for distinguishing the origins of incoming signals detected at the node since several signals may arrive at the same time if the node degree of the network is greater than one. This paper presents a signal source encoding scheme for distinguishing incoming signals efficiently, in terms of the number of detectors at each node or the number of unique wavelengths. The proposed scheme is solved by developing a new parallel genetic algorithm called distributed asynchronous genetic algorithm (DAGA). Using the DAGA, we solved signal distinction schemes for various network sizes of several topologies such as hypercube, the mesh, and the de Brujin.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.640-643
• /
• 2003

A Walsh function based associative memory is capable of storing m patterns in a single pattern storage space with Walsh encoding of each pattern. Furthermore, each stored pattern can be matched against the stored patterns extremely fast using algorithmic parallel processing. As such, this special type of memory is ideal for real-time processing of large scale information. However this incredible efficiency generates large amount of crosstalk between stored patterns that incurs mis-recognition. This crosstalk is a function of the set of different sequencies [number of zero crossings] of the Walsh function associated with each pattern to be stored. This sequency set is thus optimized in this paper to minimize mis-recognition, as well as to maximize memory saying. In this paper, this Walsh memory has been applied to the problem of face recognition, where PCA is applied to dimensionality reduction. The maximum Walsh spectral component and genetic algorithm (GA) are applied to determine the optimal Walsh function set to be associated with the data to be stored. The experimental results indicate that the proposed methods provide a novel and robust technology to achieve an error-free, real-time, and memory-saving recognition of large scale patterns.