Emerald | International Journal of Intelligent Computing and Cybernetics

ANALYTIC DESIGN OF INFORMATION GRANULATION-BASED FUZZY RADIAL BASIS FUNCTION NEURAL NETWORKS WITH THE AID OF MULTIOBJECTIVE PARTICLE SWARM OPTIMIZATION

Byoung-Jun Park, Jeoung-Nae Choi, Wook-Dong Kim, Sung-Kwun Oh — 2012-03-23 00:00:00.0

Abstract

Purpose - This study is concerned with the concept of Fuzzy Radial Basis Function Neural Networks with Information Granulation (IG- FRBFNN) and their optimization realized by means of the multiobjective Particle Swarm Optimization (MOPSO). The IG-FRBFNN uses the extended architecture of Fuzzy Radial Basis Function Neural Network (FRBFNN). In the IG-FRBFNN, we exploit the Fuzzy C-Means clustering to form the premise part of the rules. As the consequent part of fuzzy rules of the IG-based FRBFNN model (being the local model representing input output relation in the corresponding sub-space), four types of polynomials are considered, namely constant, linear, quadratic, and modified quadratic. This provides a significant level of design flexibility as each rule could come with a different type of the local model. Either the Ordinary Least Square (OLS) or the weighted Least Square (WLS) learning is exploited to estimate the values of the coefficients of the polynomial.Design/methodology/approach - In fuzzy modeling, complexity, interpretability (or simplicity) as well as accuracy of the obtained model are essential design criteria. Since the performance of the IG-RBFNN model is directly affected by some parameters such as e.g., the fuzzification coefficient used in the FCM, the number of rules and the orders of the polynomials in the consequent parts of the rules, we carry out both structural as well as parametric optimization of the network. A multi-objective Particle Swarm Optimization using Crowding Distance (MOPSO-CD) ) as well as O/WLS learning-based optimization are exploited to carry out the structural and parametric optimization of the model respectively while the optimization is of multiobjective character as it is aimed at the simultaneous minimization of complexity and maximization of accuracy. Findings - The performance of the proposed model is illustrated with the aid of three examples. The proposed optimization method leads to accurate and highly interpretable fuzzy model. Originality/value - A multi-objective Particle Swarm Optimization using Crowding Distance (MOPSO-CD)) as well as O/WLS learning-based optimization are exploited to carry out the structural and parametric optimization of the model respectively. As a result, the proposed methodology is interesting for designing accurate and highly interpretable fuzzy model.

A Version of Geiringer-like Theorem for Decision Making in the Environments with Randomness and Incomplete Information.

Boris S. Mitavskiy, Jonathan E Rowe, Chris Cannings — 2012-03-23 00:00:00.0

Abstract

Purpose - In recent years Monte-Carlo sampling methods, such as Monte Carlo tree search, have achieved tremendous success in model free reinforcement learning. A combination of the so called upper confidence bounds policy to preserve the ``exploration vs. exploitation" balance to select actions for sample evaluations together with massive computing power to store and to update dynamically a rather large pre-evaluated game tree lead to the development of software that has beaten the top human player in the game of Go on a 9 by 9 board. Much effort in the current research is devoted to widening the range of applicability of the Monte-Carlo sampling methodology to partially observable Markov decision processes with non-immediate payoffs. The main challenge introduced by randomness and incomplete information is to deal with the action evaluation at the chance nodes due to drastic differences in the possible payoffs the same action could lead to. The aim of this article is to establish a version of a theorem that originated from population genetics and has been later adopted in evolutionary computation theory that will lead to novel Monte-Carlo sampling algorithms that provably increase the AI potential. Due to space limitations the actual algorithms themselves will be presented in the sequel papers, however, the current paper provides a solid mathematical foundation for the development of such algorithms and explains why they are so promising.Design/methodology/approach - In the current paper we set up a mathematical framework, state and prove a version of a Geiringer-like theorem that is very well-suited for the development of Mote-Carlo sampling algorithms to cope with randomness and incomplete information to make decisions. From the framework it will be clear that such algorithm increase what seems like a limited sample of rollouts exponentially in size by exploiting the symmetry within the state space at little or no additional computational cost. Appropriate notions of recombination (or crossover) and schemata are introduced to stay inline with the traditional evolutionary computation terminology. The main theorem is proved using the methodology developed in the PhD thesis of the first author, however the general case of non-homologous recombination presents additional challenges that have been overcome thanks to a lovely application of the classical and elementary tool known as the ``Markov inequality" together with the lumping quotients of Markov chains techniques developed and successfully applied by the authors in the previous research for different purposes. This methodology will be mildly extended to establish the main result of the current article. In addition to establishing the Geiringer-like theorem for Monte Carlo sampling, which is the central objective of this paper, we also strengthen the applicability of the core theorem from the PhD thesis of the first author on which our main result rests. This provides additional theoretical justification for the anticipated success of the presented theory.Findings - This work establishes an important theoretical link between classical population genetics, evolutionary computation theory and model free reinforcement learning methodology. Not only the theory may explain the success of the currently existing Monte-Carlo tree sampling methodology, but it also leads to the development of novel Monte-Carlo sampling techniques guided by rigorous mathematical foundation.Practical implications - The theoretical foundations established in the current work provide guidance for the design of powerful Monte-Carlo sampling algorithms in model free reinforcement learning to tackle numerous problems in computational intelligence.Originality/value - Establishing a Geiringer-like theorem with non-homologous recombination was a long standing open problem in evolutionary computation theory. Apart from overcoming this challenge, in a mathematically elegant fashion and establishing a rather general and powerful version of the theorem, this work leads directly to the development of novel provably powerful algorithms for decision making in the environment involving randomness, hidden or incomplete information.

RLSESN based PID adaptive control for a novel wearable rehabilitation robotic hand driven by PM-TS actuators

Jun Wu, Jian Huang, Yongji Wang, Kexin Xing — 2012-03-23 00:00:00.0

Abstract

Purpose - The purpose of this paper is to develop a novel wearable rehabilitation robotic hand driven by Pneumatic Muscle - Torsion Spring (PM-TS) for finger therapy. PM has complex nonlinear dynamics, which makes PM modelling difficult. To realize high-accurate tracking for the robotic hand, an Echo State Network (ESN) based PID adaptive controller is proposed, even though the plant model is unknown.Design/methodology/approach - To drive a single joint of rehabilitation robotic hand, we propose a new PM-TS actuator comprising a Pneumatic Muscle (PM) and a Torsion Spring (TS). Based on the novel actuator, a wearable robotic hand is designed. By employing the model-free approximation capability of ESN, the RLSESN based PID adaptive controller is presented for improving the trajectory tracking performance of the rehabilitation robotic hand. An ESN together with Recursive Least Square (RLS) is called a RLSESN, where the ESN output weight matrix is updated by the online RLS learning algorithm.Findings - Practical experiments demonstrate the validity of the PM-TS actuator and indicate that the performance of the RLSESN based PID adaptive controller is better than that of the conventional PID controller. In addition, they also verify the effectiveness of the proposed rehabilitation robotic hand.Originality/value - A new PM-TS actuator configuration that uses a PM and a torsion spring for bi-directional movement of joint is presented. By utilizing the new PM-TS actuator, a novel wearable rehabilitation robotic hand for finger therapy is designed. Based on the unknown plant model, the RLSESN_PID controller is proposed to attain satisfactory performance.

Algorithm fusion to improve detection of lung cancer on chest radiographs

Gabor Horvath, Gergely Orban — 2012-03-23 00:00:00.0

Abstract

Purpose - The purpose of this paper is to show an efficient method for the detection of signs of early lung cancer. Various image processing algorithms are presented for different types of lesions, and a scheme is proposed for the combination of results.Design/methodology/approach - A computer aided detection (CAD) scheme was developed for detection of lung cancer. It enables different lesion enhancer algorithms, sensitive to specific lesion subtypes, to be used simultaneously. Three image processing algorithms are presented for the detection of small nodules, large ones, and infiltrated areas. The outputs are merged, the false detection rate is reduced with four separated support vector machine (SVM) classifiers. The classifier input comes from a feature selection algorithm selecting from various textural and geometric features. A total of 761 images were used for testing, including the database of the Japanese Society of Radiological Technology (JSRT).Findings - The fusion of algorithms reduced false positives on average by 0.6 per image, while the sensitivity remained 80%. On the JSRT database the system managed to find 60.2% of lesions at an average of 2.0 false positives per image. The effect of using different result evaluation criteria was tested and a difference as high as 4 percentage points in sensitivity was measured. The system was compared to other published methods.Originality/value - This study proves the usefulness of lesion enhancement decomposition, while proposing a scheme for the fusion of algorithms. Furthermore, a new algorithm is introduced for the detection of infiltrated areas, possible signs of lung cancer, neglected by previous solutions.

RECEDING HORIZON CONTROL FOR COOPERATIVE SEARCH OF MULTI-UAVS BASED ON DIFFERENTIAL EVOLUTION

Zhenyu Zhao, Guangshan LU — 2012-03-23 00:00:00.0

Abstract

Purpose - A hybrid method of intelligent optimization algorithm and Receding Horizon Control is presented. The method is applied to solve the problem of cooperative search of multi-UAVsDesign/methodology/approach - The intelligent optimization of DE makes the complex problem of multi-UAVs cooperative search a regular function optimization problem. To meet the real-time requirement, a thought of Receding Horizon Control is applied to. An Extended Search Map based on hormone information is used to describe the uncertain environment informationFindings - Simulation results indicate effectiveness of the hybrid method in solving the problem of cooperative search for multi-UAVsOriginality/value - The hybrid method of Differential Evolution and Receding Horizon Control is interesting for the problem of cooperative of multi-UAVs