Catálogo de publicaciones - libros

This paper proposes a recurrent wavelet-based neuro-fuzzy system (RWNFS) with the reinforcement hybrid evolutionary learning algorithm (R-HELA) for solving various control problems. The proposed R-HELA combines the compact genetic algorithm (CGA) and the modified variable-length genetic algorithm (MVGA), performs the structure/ parameter learning for dynamically constructing the RWNFS. That is, both the number of rules and the adjustment of parameters in the RWNFS are designed concurrently by the R-HELA. In the R-HELA, individuals of the same length constitute the same group. There are multiple groups in a population. The evolution of a population consists of three major operations: group reproduction using the compact genetic algorithm, variable two-part crossover, and variable two-part mutation. An illustrative example was conducted to show the performance and applicability of the proposed R-HELA method.

This paper proposes a new relation-based genetic algorithm named relational genetic algorithm (RGA) for solving partitioning problems. In our RGA, a relation-oriented representation (or relational encoding) is adopted and corresponding genetic operators are redesigned. The relational encoding is represented by the equivalence relation matrix which has a 1-1 and onto correspondence with the class of all possible partitions. It eliminates the redundancy of previous GA representations and improves the performance of genetic search. The generalized problem-independent operators we redesigned manipulate the genes without requiring specific heuristics in the process of evolution. In addition, our RGA also supports a variable number of subsets. It works without requiring a fixed number of subsets in advance. Experiments for solving some well-known classic partitioning problems by RGA and GGA with and without heuristics are performed. Experimental results show that our RGA is significantly better than GGA in all cases with larger problem sizes.

A newsboy problem model in which a vendor has limited resource is developed. It is assumed that the supplier will either sell the items to the vendor outright or offer the items to the vendor with return policy. In the latter case, the supplier buys back at certain percentage of original cost from the vendor the unsold items at the end of the selling season. The purpose of this study is to investigate how the vendor should replenish the items with return policy, constrained resources and changing procured price. Three numerical examples are provided. In one example with two constrained variables, an optimal solution is derived by using KKT (Karush-Kuhn-Tucker) conditions. The other two multiple variables examples with a minimum service level or a limited budget are solved using GA (genetic algorithm).

Fuzzy interpolative reasoning techniques can reduce the complexity of a sparse fuzzy rule-based system. In this paper, we present a new fuzzy interpola tive reasoning method via cutting and transformations techniques for sparse fuzzy rule-based systems. It produces more reasonable reasoning consequences than the ones presented in [1] and [3]. The proposed method provides a useful way to deal with fuzzy interpolative reasoning in sparse fuzzy rule-based systems.

The purpose of this study is to focus on the packaging issues at the new product introduction (NPI) stage of notebook computer that enterprises encounter when practicing global logistics. It acquires the weight of product design by quality function deployment in two phases: package design and product design. This study uses product’s attributes and their weights as the measurement index for TOPSIS to evaluate the risk priority number of FMEA. Design suggestions generated from reverse feedback can increase logistics efficiency and enable designers to design-out logistics inefficiency caused by product design at the early stage. With effective cooperation, we can learn the critical attributes in NPI when considering logistics factors and assist designers to dissolve design inefficiency. As a result, we can achieve the mechanism of prevention inefficiency in advance, decrease engineering changes, lower costs to speed up the NPI and increase enterprises’ competitiveness.

Real estate appraisal requires expert knowledge and should be performed by licensed professionals. Prior to the evaluation the appraiser must conduct a thorough study of the appraised property i.e. a land parcel and/or a building. Despite the fact that he sometimes uses the expertise of the surveyor, the builder, the economist or the mortgage lender, his estimations are usually subjective and are based on his experience and intuition. The primary goal of the paper is to present the concept of a fuzzy rule-based system to assist with real estate appraisals. The input variables of the system comprise seven attributes of a property and as the output the system proposes the property’s value. For the appraisal area, so called representative property is determined and in fact the deviations of property attribute values from the representative ones are the input into the fuzzy system. The proportion of the representative property price to the value of the property being assessed is produced as the output of the system. The experts have built the Mamdani model of the system, however they have not been able to construct the rule base. Therefore an evolutionary algorithm has been employed to generate the rule base. The Pittsburgh approach has been applied. The learning process has been conducted using training and testing sets prepared on the basis of 150 sales transactions from one city.

This study presents a fuzzy system by collecting membership function and rules based on a decision model that uses empirical Bayesian estimation to construct a server-dependent / /2 / queue. A Markovian queue of finite capacity in which the number of servers depends upon queue length is considered. First, data on the interarrival times and service times are collected by observing a queuing system, and the empirical Bayesian method is adopted to estimate its server utilization. Second, the costs are associated with the operation of the second server and the waiting of customers, to establish a cost minimization model to determine the optimal number of customers in the system to activate the second server (), and the optimal number of customers in system to deactivate the second server (). The decision model provides knowledge to construct rules used in a fuzzy inference system. The MATLAB Fuzzy Inference Toolbox is used to construct a fuzzy system to aid management in determining when to initiate the second server and when to turn it off, according to specific parameters.

This paper presents techniques that enable a talker tracking for effective human-robot interaction. We propose new way of integrating an EM algorithm and a particle filter to select an appropriate path for tracking the talker. It can easily adapt to new kinds of information for tracking the talker with our system. This is because our system estimates the position of the desired talker through means, variances, and weights calculated from EM training regardless of the numbers or kinds of information. In addition, to enhance a robot’s ability to track a talker in real-world environments, we applied the particle filter to talker tracking after executing the EM algorithm. We also integrated a variety of auditory and visual information regarding sound localization, face localization, and the detection of lip movement. Moreover, we applied a sound classification function that allows our system to distinguish between voice, music, or noise. We also developed a vision module that can locate moving objects.

The traditional approach to implement motor behaviour in a robot required a programmer to carefully decide the joint velocities at each timestep. By using the principle of learning by imitation, the robot can instead be taught simply by it what to do. This paper investigates the self-organization of a connectionist modular architecture for motor learning and control that is used to imitate human dancing. We have observed that the internal representation of a motion behaviour tends to be captured by more than one module. This supports the hypothesis that a modular architecture for motor learning is capable of self-organizing the decomposition of a movement.

In this paper, a new hybrid particle swarm optimization model named HPSO that combines random-key (RK) encoding scheme, individual enhancement (IE) scheme, and particle swarm optimization (PSO) is presented and used to solve the flow-shop scheduling problem (FSSP). The objective of FSSP is to find an appropriate sequence of jobs in order to minimize makespan. Makespan means the maximum completion time of a sequence of jobs running on the same machines in flow-shops. By the RK encoding scheme, we can exploit the global search ability of PSO thoroughly. By the IE scheme, we can enhance the local search ability of particles. The experimental results show that the solution quality of FSSP based on the proposed HPSO is far better than those based on GA [1] and NPSO [1], respectively.