Applied Intelligence

Zhu, Bing; He, Changzheng; Liatsis, Panos

doi: 10.1007/s10489-010-0244-1pmid: N/A

Missing data imputation is an important research topic in data mining. The impact of noise is seldom considered in previous works while real-world data often contain much noise. In this paper, we systematically investigate the impact of noise on imputation methods and propose a new imputation approach by introducing the mechanism of Group Method of Data Handling (GMDH) to deal with incomplete data with noise. The performance of four commonly used imputation methods is compared with ours, called RIBG (robust imputation based on GMDH), on nine benchmark datasets. The experimental result demonstrates that noise has a great impact on the effectiveness of imputation techniques and our method RIBG is more robust to noise than the other four imputation methods used as benchmark.

Kang, Jeong-Gwan; Kim, Sunhyo; An, Su-Yong; Oh, Se-Young

doi: 10.1007/s10489-010-0257-9pmid: N/A

This paper presents Neuro-Evolutionary Optimization SLAM (NeoSLAM) a novel approach to SLAM that uses a neural network (NN) to autonomously learn both a nonlinear motion model and the noise statistics of measurement data. The NN is trained using evolutionary optimization to learn the residual error of the motion model, which is then added to the odometry data to obtain the full motion model estimate. Stochastic optimization is used, to accommodate any kind of cost function. Prediction and correction are performed simultaneously within our neural framework, which implicitly integrates the motion and sensor models. An evolutionary programming (EP) algorithm is used to progressively refine the neural model until it generates a trajectory that is most consistent with the actual sensor measurements. During this learning process, NeoSLAM does not require any prior knowledge of motion or sensor models and shows consistently good performance regardless of the robot and the sensor noise type. Furthermore, NeoSLAM does not require the data association step at loop closing which is crucial in most other SLAM algorithms, but can still generate an accurate map. Experiments in various complex environments with widely-varying types of noise show that the learning capability of NeoSLAM ensures performance that is consistently less sensitive to noise and more accurate than that of other SLAM methods.

Xu, Yong; Ohmoto, Yoshimasa; Okada, Shogo; Ueda, Kazuhiro; Komatsu, Takanori; Okadome, Takeshi; Kamei, Koji; Sumi, Yasuyuki; Nishida, Toyoaki

doi: 10.1007/s10489-010-0255-ypmid: N/A

When an adaptive agent works with a human user in a collaborative task, in order to enable flexible instructions to be issued by ordinary people, it is believed that a mutual adaptation phenomenon can enable the agent to handle flexible mapping relations between the human user’s instructions and the agent’s actions. To elucidate the conditions required to induce the mutual adaptation phenomenon, we designed an appropriate experimental environment called “WAITER” (Waiter Agent Interactive Training Experimental Restaurant) and conducted two experiments in this environment. The experimental results suggest that the proposed conditions can induce the mutual adaptation phenomenon.

Xu, Ying; Qu, Rong

doi: 10.1007/s10489-010-0256-xpmid: N/A

This paper investigates the first hybrid scatter search and path relinking meta-heuristic for the Delay-Constrained Least-Cost (DCLC) multicast routing problem. The underpinning mathematic model of the DCLC multicast routing problem is the constrained Steiner tree problem in graphs, a well known NP-complete problem. After combining a path relinking method as the solution combination method in scatter search, we further explore two improvement strategies: tabu search and variable neighborhood search, to intensify the search in the hybrid scatter search algorithm. A large number of simulations on some benchmark instances from the OR-library and a group of random graphs of different characteristics demonstrate that the improvement strategy greatly affects the performance of the proposed scatter search algorithm. The hybrid scatter search algorithm intensified by a variable neighborhood descent search is highly efficient in solving the DCLC multicast routing problem in comparison with other algorithms and heuristics in the literature.

Kuzu, Mehmet; Cicekli, Nihan

doi: 10.1007/s10489-010-0238-zpmid: N/A

In this paper, novel ideas are presented for solving the automated web service composition problem. Some of the possible real world problems such as partial observability of the environment, nondeterministic effects of web services and service execution failures are solved through a dynamic planning approach. The proposed approach is based on a novel AI planner that is designed for working in highly dynamic environments under time constraints, namely Simplanner. World altering service calls are done according to the WS-Coordination and WS-Business Activity web service transaction specifications in order to physically recover from failure situations and prevent the undesired side effects of the aborted web service composition efforts.

Mazinan, A.; Sheikhan, M.

doi: 10.1007/s10489-010-0253-0pmid: N/A

This paper describes a novel artificial intelligence based predictive control scheme for the purpose of dealing with so many complicated systems. In the control scheme proposed here, the system has to be first represented through a multi-Takagi-Sugeno-Kang (TSK) fuzzy-based model approach to make an appropriate prediction of the system behavior. Subsequently, a multi-generalized predictive control (GPC) scheme, which is organized based on a number of GPC schemes, is realized in line with the investigated model outcomes, at chosen operating points of the system. In case of the proposed control strategy realization, the investigated multi-GPC scheme is instantly updated to handle the system by activating the best control scheme through a new GPC identifier, while the system output is suddenly varied with respect to time. To present the applicability of the proposed control scheme, an industrial tubular heat exchanger system and also a drum-type boiler-turbine system have been chosen to drive through the proposed strategy. In such a case, the simulations are carried out and the corresponding results are compared with those obtained using traditional GPC scheme in addition to nonlinear GPC (NLGPC) scheme, as benchmark approaches, where the acquired results of the proposed control scheme are desirably verified.

Zhang, Shichao; Chen, Feng; Wu, Xindong; Zhang, Chengqi; Wang, Ruili

doi: 10.1007/s10489-010-0247-ypmid: N/A

Bridging rules take the antecedent and action from different conceptual clusters. They are distinguished from association rules (frequent itemsets) because (1) they can be generated by the infrequent itemsets that are pruned in association rule mining, and (2) they are measured by their importance including the distance between two conceptual clusters, whereas frequent itemsets are measured only by their support. In this paper, we first design two algorithms for mining bridging rules between clusters, and then propose two non-linear metrics to measure their interestingness. We evaluate these algorithms experimentally and demonstrate that our approach is promising.

Zhang, De-gan

doi: 10.1007/s10489-010-0245-0pmid: N/A

Seamless migration is one of pervasive computing applications. The function of seamless mobility is suitable for mobile services such as mobile Web-based learning. In this paper, we propose an approach that supports an attentive mobile learning paradigm. This mobile learning dynamically follows the user from place to place and machine to machine without user’s awareness or intervention by active service. This capability can be obtained by component-based smart system and agent-based migrating mechanism. To demonstrate the approach, the theoretical background of fuzzy-neural network for attentive service will be explained. The proposed fusion decision method is based on fuzzy-neural network which can make the input signal data to fuse better. Using online tuning, the fusion processing can be accelerated and the fusion belief degree can be improved. Description of mobile learning task and migrating granularity of the task is suggested. The design of the seamless migration mechanism is introduced. This includes solving several important sub-problems, such as transferring delay, transferring failure, and residual computation dependency. Our implemented system for attentive mobile learning based on seamless migration is presented. The validity comparison and evaluation of this kind of mobile learning paradigm is shown by experimental demos. This suggested attentive mobile learning paradigm based on seamless migration is useful and convenient to mobile learners.

Altınçay, Hakan; Erenel, Zafer

doi: 10.1007/s10489-010-0250-3pmid: N/A

In this study, the differences among widely used weighting schemes are studied by means of ordering terms according to their discriminative abilities using a recently developed framework which expresses term weights in terms of the ratio and absolute difference of term occurrence probabilities. Having observed that the ordering of terms is dependent on the weighting scheme under concern, it is emphasized that this can be explained by the way different schemes use term occurrence differences in generating term weights. Then, it is proposed that the relevance frequency which is shown to provide the best scores on several datasets can be improved by taking into account the way absolute difference values are used in other widely used schemes. Experimental results on two different datasets have shown that improved F 1 scores can be achieved.

Soak, Sang-Moon; Lee, Sang-Wook

doi: 10.1007/s10489-010-0248-xpmid: N/A

In this paper, we propose a new memetic algorithm for the quadratic multiple container packing problem. The proposed memetic algorithm is based on the adaptive link adjustment evolutionary algorithm (ALA-EA) and it incorporates heuristic fitness improvement schemes into the ALA-EA. In addition, we propose a new powerful initialization method for the QMCPP.