Hybrid self-inertia weight adaptive particle swarm optimisation with local search using C4.5 decision tree classifier for feature selection problems

ABSTRACT

Feature selection is an important task to improve the classifier’s accuracy and to decrease the problem size. A number of methodologies have been presented for feature selection problems using metaheuristic algorithms. In this paper, an improved self-adaptive inertia weight particle swarm optimisation with local search and combined with C4.5 classifiers for feature selection algorithm is proposed. In this proposed algorithm, the gradient base local search with its capacity of helping to explore the feature space and an improved self-adaptive inertia weight particle swarm optimisation with its ability to converge a best global solution in the search space. Experimental results have verified that the SIW-APSO-LS performed well compared with other state of art feature selection techniques on a suit of 16 standard data sets.     

A Novel Beam Search to Improve Neural Machine Translation for English-Chinese

Neural Machine Translation (NMT) is an end-to-end learning approach for automated translation, overcoming the weaknesses of conventional phrase-based translation systems. Although NMT based systems have gained their popularity in commercial translation applications, there is still plenty of room for improvement. Being the most popular search algorithm in NMT, beam search is vital to the translation result. However, traditional beam search can produce duplicate or missing translation due to its target sequence selection strategy. Aiming to alleviate this problem, this paper proposed neural machine translation improvements based on a novel beam search evaluation function. And we use reinforcement learning to train a translation evaluation system to select better candidate words for generating translations. In the experiments, we conducted extensive experiments to evaluate our methods. CASIA corpus and the 1,000,000 pairs of bilingual corpora of NiuTrans are used in our experiments. The experiment results prove that the proposed methods can effectively improve the English to Chinese translation quality.     

A disruption recovery plan in a three-stage production-inventory system

This paper proposes a recovery plan for managing disruptions in a three-stage production-inventory system under a mixed production environment. First, a mathematical model is developed to deal with a disruption at any stage while maximizing total profit during the recovery-time window. The model is solved after the occurrence of a disruption event, with changed data used to generate a revised plan. We also propose a new and efficient heuristic for solving the developed mathematical model. Second, multiple disruptions are considered, where a new disruption may or may not affect the recovery plans of earlier disruptions. The heuristic, developed for a single disruption, is extended to deal with a series of disruptions so that it can be implemented for disruption recovery on a real-time basis. We compare the heuristic solutions with those obtained by a standard search algorithm for a set of randomly generated disruption test problems, and that show the consistent performance of our developed heuristic with lower computational times. Finally, some numerical examples and a real-world case study are presented to demonstrate the benefits and usefulness of our proposed approach.     

Using provenance to efficiently improve metadata searching performance in storage systems

In cloud storage systems, more than 50% of requests are metadata operations, and thus the file system metadata search performance has become increasingly important to different users. With the rapid growth of storage system scales in volume, traditional full-size index trees cannot offer high-performance metadata search due to hierarchical indexing bottleneck. In order to alleviate the long latency and improve the quality-of-service (QoS) in cloud storage service, we proposed a novel provenance based metadata-search system, called PROMES. The metadata search in PROMES is split into three phases: (i) leveraging correlation-aware metadata index tree to identify several files as seeds, most of which can satisfy the query requests, (ii) using the seeds to find the remaining query results via relationship graph search, (iii) continuing to refine and rerank the whole search results, and sending the final results to users. PROMES has the salient features of high query accuracy and low latency, due to files’ tight and lightweight indexing in relationship graph coming from provenance’s analysis. Compared with state-of-the-art metadata searching schemes, PROMES demonstrates its efficiency and efficacy in terms of query accuracy and response latency.     

Dynamic design of sales territories

We introduce the Multiple Traveling Salesmen and Districting Problem with Multi-periods and Multi-depots. In this problem, the compactness of the subdistricts, the dissimilarity measure of districts and an equity measure of salesmen profit are considered as part of the objective function, and the salesman travel cost on each subdistrict is approximated by the Beardwood–Halton–Hammersley formula. An adaptive large neighbourhood search metaheuristic is developed for the problem. It was tested on modified Solomon and Gehring & Homberger instances. Computational results confirm the effectiveness of the proposed metaheuristic.     

Promising approaches to improve the performances of hybrid non‐isocyanate polyurethane

The performances of hybrid non‐isocyanate polyurethane (NIPU) materials were improved by using promising approaches. For the ring‐opening reaction of cyclic carbonate with excess amine, methanol was selected as solvent to increase the molar mass of NH₂‐terminated NIPU pre‐polymer but also caused the formation of urea as a result of eliminating hydrogen bonding between NIPU chains. Meanwhile, another approach was carried out to reduce the content of urea by altering the introduction mode of reactants, along with further increasing the molar mass of the NIPU pre‐polymer. The final hybrid materials were prepared by adding bisphenol‐A diglycidyl ether into the pre‐polymer. The results indicated that the introduction of methanol and the change of the introduction mode of reactants could improve the performance of the hybrid materials. The hybrid material derived from the forward method has optimal performance with a tensile strength of 10.8 MPa, elongation at break of 167%, a glass transition temperature of 49.1 °C and a crosslinking density of 450 mol m^?3 as well as presentable thermodynamic stability, ascribed to the high molar mass of the NIPU pre‐polymer, the low urea content and the dense crosslinking network. © 2018 Society of Chemical Industry     

变邻域量子烟花算法求解CVRP

针对带容量约束的车辆路径问题，提出一种融合量子进化算法和变邻域优化策略的变邻域量子烟花算法。该算法采用等分随机键与最大位置法结合的实数编码方式，通过量子旋转门和非门变异提高算法全局搜索能力，同时运用结合2-Opt的变邻域优化策略加强局部搜索能力。选取17个基准算例进行参数实验和对比实验，实验结果表明，相对于对比算法，所提出的算法具有较好的寻优能力和收敛速度。     

航班恢复问题的迭代局部搜索算法

在恶劣天气和机械故障等原因造成航班不能按照原计划执行时，航空公司需要采取相应的措施对航班进行恢复。本文基于经典的资源指派模型，综合考虑了调整时间、换机、联程拉直、取消航班和调机5种恢复策略，提出一种以最小化加权成本为优化目标的航班恢复模型，并设计一种迭代局部搜索算法。首先用构造-修复启发式方法构造可行解，然后从该初始解出发，在飞机路线对的邻域中进行局部搜索。当陷入局部最优后，对解进行扰动，然后从扰动后的解重新出发进行局部搜索。为了提高搜索效率，同时降低陷入局部最优解的概率，局部搜索过程采用模拟退火算法。实例结果表明，本文提出的模型及算法能够在短时间内对受到影响的大规模航班计划进行恢复。     

Improved non-maximum suppression for object detection using harmony search algorithm

Non-maximum suppression (NMS) plays a key role in many modern object detectors. It is responsible to remove detection boxes that cover the same object. NMS greedily selects the detection box with maximum score; other detection boxes are suppressed when the degree of overlap between these detection boxes and the selected box exceeds a predefined threshold. Such a strategy easily retain some false positives, and it limits the ability of NMS to perceive nearby objects in cluttered scenes. This paper proposes an effective method combining harmony search algorithm and NMS to alleviate this problem. This method regards the task of NMS as a combination optimization problem. It seeks final detection boxes under the guidance of an objective function. NMS is applied to each harmony to remove imprecise detection boxes, and the remaining boxes are used to calculate the fitness value. The remaining detection boxes in a harmony with highest fitness value are chosen as the final detection results. The standard Pattern Analysis, Statistical Modeling and Computational Learning Visual Object Classes dataset and the Microsoft Common Objects in Context dataset are used in all of the experiments. The proposed method is applied to two popular detection networks, namely Faster Region-based Convolutional Neural Networks and Region-based Fully Convolutional Networks. The experimental results show that the proposed method improves the average precision of these two detection networks. Moreover, the location performance and average recall of these two detectors are also improved.     

Backtracking search optimization heuristics for nonlinear Hammerstein controlled auto regressive auto regressive systems

In this work, novel application of evolutionary computational heuristics is presented for parameter identification problem of nonlinear Hammerstein controlled auto regressive auto regressive (NHCARAR) systems through global search competency of backtracking search algorithm (BSA), differential evolution (DE) and genetic algorithms (GAs). The mean squared error metric is used for the fitness function of NHCARAR system based on difference between actual and approximated design variables. Optimization of the cost function is conducted with BSA for NHCARAR model by varying degrees of freedom and noise variances. To verify and validate the worth of the presented scheme, comparative studies are carried out with its counterparts DE and GAs through statistical observations by means of weight deviation factor, root of mean squared error, and Thiel’s inequality coefficient as well as complexity measures.