International Journal of Computational Intelligence and Applications

Feature ordering is a significant data preprocessing method in incremental attribute learning (IAL), where features are gradually trained according to a given order. Previous research showed feature ordering is crucial to the IAL performance. It is relevant to each feature's discrimination ability, which can be calculated by single discriminability (SD). However, when feature dimensions increase, feature discrimination ability should also be calculated incrementally, because discrimination ability in lower dimensional spaces is different from that in higher spaces. Thus based on SD, accumulative discriminability (AD), a new statistical metric for incremental feature discrimination ability estimation, is designed. Moreover, a criterion that summarizes all the produced values of AD is employed to obtain the optimum feature ordering for classification problems based on neural networks by means of IAL. In addition, in order to reduce the time consumption, an effective feature ordering approach is developed. Compared with the feature ordering obtained by other approaches, the method outlined in this paper obtained good final classification results, which indicates that, firstly, feature discrimination ability should be incrementally estimated in IAL; and secondly, feature ordering derived by AD and its corresponding approaches are applicable with IAL.

Type-2 fuzzy sets are generalizations of ordinary fuzzy sets, in which membership grades are characterized by fuzzy membership functions. Here, a problem of finding distance between two interval type-2 fuzzy sets (IT2-FSs) was considered. Based on a new definition of centroid for an IT2-FS, a formulation for calculation of the distance between two IT2-FSs was introduced, and an algorithm was explained to obtain it. The proposed distance formula was incorporated in Yang and Shih's clustering algorithm to reach a clustering method for interval type-2 fuzzy data sets. The applicability of the proposed distance formula was evaluated using two artificial and real data sets, and reasonable results were obtained.

Petroleum geomechanics characterization refers to the process of quantitatively assigning geomechanical parameters using all available field data. In this paper, an attempt is made to develop a computational intelligence method that integrates genetic algorithm (GA) and artificial neural network (ANN). Through this method, these geomechanical parameters such as horizontal in situ stresses, fracture aperture and joint spacing are determined based on the borehole displacements during drilling well. In the hybrid ANN–GA model, GA can automatically identify geomechanical parameters as neural inputs from monitored borehole displacements, and the ANN is trained to learn the complex relationship between geomechanical parameters and target borehole displacements. Data from numerical experiments on petroleum wells are used in verification of the proposed computational intelligence method for geomechanics characterization. The study of numerical experiment illustrates that the proposed computational intelligence method has the ability to generate reliable results.

Construction procurement is a key business where price negotiation is commonly required to reach final contractual agreement. However, even simple negotiations often result in infeasible agreements. The uncertain and limited supplier information as well as complex correlations among various factors affecting supplier behaviors make the contractor difficult to decide the appropriate offer price (OP) and vice versa. This study proposes a novel Fuzzy Bayesian Game Model (FBGM) for improving the prediction effectiveness of negotiation behaviors. The performance of the proposed FBGM was evaluated in the case where an agent uses the counter-OP of an opponent to learn the negotiation strategy of the opponent. The validation analysis shows that the sequential updating process of FBGM significantly improves the estimation ability of negotiators. The proposed model also gives a comprehensive view of negotiation scenarios by considering all possible negotiation cases. Using FBGM, negotiators can apply flexible strategies to optimize their own profit with a reasonable negotiation time.

The main purpose of this study is to employ Pi-Sigma Neural Network (PSNN) for one-step-ahead temperature forecasting. In this paper, we evaluate the performances of PSNN by comparing the network model with widely used Multilayer Perceptron (MLP). PSNN which is a class of Higher Order Neural Networks (HONN), has a highly regular structure, needs much smaller number of weights and less training time. The PSNN is use to overcome the drawbacks of MLP, which can easily trapped into local minima and prone to overfit. Both network models were trained with standard backpropagation algorithm. Through 1012 experiments, it has been demonstrated that the PSNN has a high practicability and better temperature forecasting for one-step-ahead using historical temperature data of Batu Pahat region.

In recent studies we found that there are many optimization methods presented for multicore processor performance optimization, however each method is suffered from limitations. Hence in this paper we presented a new method which is a combination of bacterial Foraging Particle swarm Optimization with certain constraints named as Constraint based Bacterial Foraging Particle Swarm Optimization (CBFPSO) scheduling can be effectively implemented. The proposed Constraint based Bacterial Foraging Particle Swarm Optimization (CBFPSO) scheduling for multicore architecture, which updates the velocity and position by two bacterial behaviours, i.e. reproduction and elimination dispersal. The performance of CBFPSO is compared with the simulation results of GA, and the result shows that the proposed algorithm has pretty good performance on almost all types of cores compared to GA with respect to completion time and energy consumption.

Dynamic routing problems ask for policies and algorithms to build reactive plans able to integrate dynamically generated requests in currently running plans. The main desired quality of such approaches is to overcome the lack of information availability in the beginning of the solving process. That is by proposing flexible partial solutions anticipating future requests. In this paper, we propose a neural model based on Hierarchical self-organizing maps (HSOM) for solving the dynamic vehicle routing problem (DVRP). The DVRP consists of finding routes for a set of vehicles to serve customer's dynamically issued requests while minimizing routing costs and satisfying some constraints. We present the concept of problem control bloc (PCB) to represent the problem to solve by the HSOM algorithm when triggered by a new event. The overall solving approach is integrated within a discrete event manager that will wait for any new event to build the PCB, call the HSOM solver, get the proposed solution and update the running plan accordingly. The proposed approach was applied to solve the DVRP problems adapted from VRP benchmarks. The obtained results are compared to the best known solutions.