This paper considers a facility construction problem in a rectangular urban area with some barriers and rectilinear distance. There exist some demand points and possible construction sites with preference of construction. For each site, its construction cost is a fuzzy random variable due to the change of the situation and difficulty of estimation after planning. Therefore in this case the total cost is also a fuzzy random variable and so probability that the cost becomes below the budget should not be below the fixed level from randomness of the total cost. We should decide suitable facility construction places under the condition that each demand point is covered within some distance from at least one facility. Under this condition, the possibility that the above chance constraint is satisfied should be maximized and minimal preference among facility constructed sites maximized. We seek some non-dominated solutions after definition of non-domination.
This paper proposes the model to obtain group decision as the approximation of the individual decisions. Both the group and individual decisions are denoted as the interval weights of alternatives by Interval AHP. The deviations between the individual and group decisions are measured based on the satisfaction and dissatisfaction of each decision maker. S/he is satisfied with the group decision, when his/her decision is included in it. While, s/he is dissatisfied with it, when his/hers does not include it. Then, the group decision is obtained so as to be common to all decision makers' decisions and to maximize and minimize their satisfactions and dissatisfactions, respectively.
Optimal design problems of IIR (Infinite Impulse Response) filters with finite word length are formulated as nonlinear discrete optimization problems, and as the filter order and the word length of coefficients increase, the calculation cost for solving this problem rises exponentially. Moreover, since the guarantee of stability is necessary in the IIR filter design, the additional calculation cost of this must also be considered. For solving these problems, we express the transfer functions of filters as the product (cascade connection) of second order rational functions to make the guarantee of stability easily, and propose an efficient solution method based on PSO for optimal design problems of such filters which can obtain an (approximate) optimal solution in practical time for high-order problems.
We discuss modeling methods for a multi objective problem with fuzzy constraints. First, we explain the composition adjustment problem of the molten steel in a steel making process to clarify the features of the problem. Next, the limitations of the conventional goal programming method are shown. Further, we describe the effectiveness of the fuzzy goal programming method using some examples.
This paper present Adaptive Mapping Networks (AMNs) as an adaptive and incremental method to learn series data for visualizing on a category map. The architecture of AMNs comprises three modules: codebook module, labeling module, and mapping module. The codebook module quantizes input data as codebooks of low-dimensional feature vectors using Self-Organizing Maps. The labeling module creates labels as a candidate of categories based on the incremental and adaptive learning of Adaptive Resonance Theory. The mapping modle visualizes spatial relations of categories on a category map using Counter Propagation Networks. AMNs actualize supervised learning and unsupervised learning to change its network structures. The experimental results using open datasets of two types show the recognition accuracy of our method is superior than that of the existing method. Moreover we present the usefulness of visualizing functions using category maps.
In this paper, we integrated Discounted UCB1-tuned, which uses weighted value and weighted variance, into Q-learning agents and experimentally evaluated its performance. Discounted UCB1-tuned is an optimized selection method that balances exploration and exploitation and outperforms other methods, including ε-greedy. We conducted experiments on the effect of default values and learning rate in a multi-armed bandit problem. Our algorithm selects actions its value is not updated or with the highest UCB value in updatable state-actions. We show the results of the continuous state spaces shortest path problem followed by a discussion.