冷启动下用户特征模型的构建与个性化推荐

针对个性化推荐中的冷启动和用户模型主观个性特征描述不足的问题,提出一种基于用户初始特征模型优化构建的个性化推荐方法.通过对成对比较矩阵构建方法的优化和改进,减少提取主观性权重比较结果时,用户的比较操作次数,通过推导计算得出用户的初始特征模型,并据此完成推荐.通过将该方法应用到周边美食个性化推荐中,验证该方法所建立的初始...     

Addressing the production balancing problem

The problem of achieving a uniform flow rate through a production line and controlling in-process inventories, i.e. balancing a line, has been the subject of prior research. Much of this research has been directed at identifying “optimal” solutions to the problem. The implementation of derived techniques, however, has been ignored. Particularly, some of the behavioral factors involved with implementing a technique has been disregarded. This paper describes an approach which addresses the “balancing problem” but considers the problem of implementation. The approach specifically examines production personnel reassignments as a key implementation factor.Some prior research studies identify variable operator performance times as a key factor relating to less than uniform flow rates and excess inventories. The approach proposed in this study demonstrates that average operator performance times can be used when multiple periods are used to implement a desired solution.     

Aggregate location recommendation in dynamic transportation networks

Travel planning and location recommendation are increasingly important in recent years. In this light, we propose and study a novel aggregate location recommendation query (ALRQ) of discovering aggregate locations for multiple travelers and planning the corresponding travel routes in dynamic transportation networks. Assuming the scenario that multiple travelers target the same destination, given a set of travelers’ locations Q, a set of potential aggregate location O, and a departure time t, the ALRQ finds an aggregate location o ∈ O that has the minimum global travel time \({\sum }_{q \in Q} T(q,o,t)\), where T(q,o,t) is the travel time between o and q with departure time t. The ALRQ problem is challenging due to three reasons: (1) how to model the dynamic transportation networks practically, and (2) how to compute ALRQ efficiently. We take two types of dynamic transportation networks into account, and we define a pair of upper and lower bounds to prune the search space effectively. Moreover, a heuristic scheduling strategy is adopted to schedule multiple query sources. Finally, we conducted extensive experiments on real and synthetic spatial data to verify the performance of the developed algorithms.     

Solving the uncapacitated hub location problem using genetic algorithms

Hub location problems are widely studied in the area of location theory, where they involve locating the hub facilities and designing the hub networks. In this paper, we present a new and robust solution based on a genetic search framework for the uncapacitated single allocation hub location problem (USAHLP). To present its effectiveness, we compare the solutions of our GA-based method with the best solutions presented in the literature by considering various problem sizes of the CAB data set and the AP data set. The experimental work demonstrates that even for larger problems the results of our method significantly surpass those of the related work with respect to both solution quality and the CPU time to obtain a solution. Specifically, the results from our method match the optimal solutions found in the literature for all test cases generated from the CAB data set with significantly less running time than the related work. For the AP data set, our solutions match the best solutions of the reference study with an average of 8 times less running time than the reference study. Its performance, robustness and substantially low computational effort justify the potential of our method for solving larger problem sizes.     

Solving the uncapacitated facility location problem using tabu search

A tabu search heuristic procedure is developed to solve the uncapacitated facility location problem. Tabu search is used to guide the solution process when evolving from one solution to another. A move is defined to be the opening or closing of a facility. The net cost change resulting from a candidate move is used to measure the attractiveness of the move. After a move is made, the net cost change of a candidate move is updated from its old value. Updating, rather than re-computing, the net cost changes substantially reduces computation time needed to solve a problem when the problem is not excessively large. Searching only a small subset of the feasible solutions that contains the optimal solution, the procedure is computationally very efficient. A computational experiment is conducted to test the performance of the procedure and computational results are reported. The procedure can easily find optimal or near optimal solutions for benchmark test problems from the literature. For randomly generated test problems, this tabu search procedure not only obtained solutions completely dominating those obtained with other heuristic methods recently published in the literature but also used substantially less computation time. Therefore, this tabu search procedure has advantage over other heuristic methods in both solution quality and computation speed.     

融合地点影响力的兴趣点推荐算法

为解决兴趣点（POI）推荐不准确和效率低的问题,深入分析社交因素和地理位置因素的影响,提出了一种融合地点影响力的POI推荐算法。首先,为了解决签到数据稀疏的问题,将2-度好友引入协同过滤算法中构建了社交影响模型,通过计算经历和好友相似度获取2-度好友对用户的社交影响;其次,深入考虑地理位置因素对POI推荐影响,在对社交网络分析的基础上构造了地点影响力模型,通过PageRank算法发现用户影响力,结合POI被签到次数计算地点影响力,获取准确的整体位置偏好,并使用核密度估计方法对用户签到行为建模和获取个性化地理位置特征;最后,融合社交模型和地理位置模型提高推荐准确性,并通过构造POI推荐候选集来提高推荐效率。在Gowalla和Yelp签到数据集上实验,结果表明所提算法能够快速完成POI推荐,在准确率和召回率指标上明显优于融合时间因素的位置推荐（LRT）和融合地理社交因素的个性化位置推荐（iGSLR）算法。     

Noise-tolerance matrix completion for location recommendation

Due to the sharply increasing number of users and venues in Location-Based Social Networks, it becomes a big challenge to provide recommendations which match users’ preferences. Furthermore, the sparse data and skew distribution (i.e., structural noise) also worsen the coverage and accuracy of recommendations. This problem is prevalent in traditional recommender methods since they assume that the collected data truly reflect users’ preferences. To overcome the limitation of current recommenders, it is imperative to explore an effective strategy, which can accurately provide recommendations while tolerating the structural noise. However, few study concentrates on the process of noisy data in the recommender system, even recent matrix-completion algorithms. In this paper, we cast the location recommendation as a mathematical matrix-completion problem and propose a robust algorithm named Linearized Bregman Iteration for Matrix Completion (LBIMC), which can effectively recover the user-location matrix considering structural noise and provide recommendations based solely on check-in records. Our experiments are conducted by an amount of check-in data from Foursquare, and the results demonstrate the effectiveness of LBIMC.     

基于位置社交网络的地点推荐算法

目前基于协同过滤的地点推荐算法存在难以准确估算用户偏好、推荐结果准确性不高等问题。改进了传统协同过滤中相似用户计算方法,在迭代过程中分别计算用户相似度和地点相似度的值,并不断交叉调整对方的值,直至收敛。该方法能够在稀疏的数据集下准确计算用户相似性。此外,在top-N推荐阶段,同时考虑了用户的兴趣度和推荐地点离用户所在地距离的影响,并设置阈值控制二者的权重,自适应地产生推荐结果。实验表明,与其它方法相比该方法能够获得更好的推荐效果。     

Addressing image and Poisson noise deconvolution problem using deep learning approaches

Digital images are more important in numerous contemporary applications, and the need for images in the technical field is also increasing drastically. It is used to recognize signatures and faces in many industries and is applicable for intelligent departments. The images are usually associated with the noise content; this may happen due to the instrument imperfections, troubleshooting while collecting data from the acquisition process, and another natural phenomenon. Poisson noise, also known as photon noise, is caused in the images due to the statistical essence of electromagnetic waves. X-ray, visible light, and gamma rays are electromagnetic waves. The enhancement of the convolution model in addressing images is challenging due to the various constituents such as optical aberrations, noise level, and optical setup. The modeling configuration of the image is attained using the point spread function (PSF), which is responsible for the system's impulse response. The quality image is retrieved by denoising and super-resolution (SR) methods; these methods simultaneously eliminate the noise content from the images. A Richardson–Lucy and alternating direction method of multipliers type of non-blind iterative algorithmic approaches associated with the PSF performance in addressing image is comparatively analyzed. The deep learning approach, convolutional neural networks (CNNs), is also employed to understand the nonlinear mapping relationship between the observed data and ground reality. The performance of the various network approaches is compared in this article. The result obtained shows that the deep learning CNNs achieved higher accuracy in producing denoising images. The goal of the proposed system model is to remove the interference noise in images. The high-resolution images are obtained by implementing a SR-based CNN model.     

基于时间感知的地点推荐算法

基于位置社交网络的服务层出不穷,而地点推荐系统是其最主要的应用之一。目前地点推荐算法均致力于提高用户对地点的兴趣度预测,没有考虑时间因素对推荐结果的影响。事实上人们是否访问某一地点,与其所处的时间是紧密相关的。因此提出了一种概率模型,在统一的框架下将用户的兴趣度、用户所处时间和地点自身的流行度三个因素综合考虑,并在真实的数据集Foursquare上进行了测试。实验表明,与其他的方法相比,本方法能够获得更好的推荐效果,增强了用户体验。     

基于仿射传播聚类的大规模选址布局问题求解*

选址布局问题是一个广泛研究的运筹学问题。通过将选址布局问题看做聚类问题,结合仿射传播聚类算法并且将候选地址的信息映射为特征向量,提出了两种求解选址布局问题的方法：基于块划分的选址布局方法和基于道路网络的选址布局方法。使用模拟数据集和真实数据集来评估两种方法,实验结果表明两种方法都能够求解设施资源受限和不受限情况下的选址布局问题,而且可以很好地解决大规模的选址布局问题。     

Addressing a nonlinear fixed-charge transportation problem using a spanning tree-based genetic algorithm

In this paper, we consider the fixed-charge transportation problem (FCTP) in which a fixed cost, sometimes called a setup cost, is incurred if another related variable assumes a nonzero value. To tackle such an NP-hard problem, there are several genetic algorithms based on spanning tree and Prüfer number representation. Contrary to the findings in previous works, considering the genetic algorithm (GA) based on spanning tree, we present a pioneer method to design a chromosome that does not need a repairing procedure for feasibility, i.e. all the produced chromosomes are feasible. Also, we correct the procedure provided in previous works, which designs transportation tree with feasible chromosomes. We show the previous procedure does not produce any transportation tree in some situations. Besides, some new crossover and mutation operators are developed and used in this work. Due to the significant role of crossover and mutation operators on the algorithm’s quality, the operators and parameters need to be accurately calibrated to ensure the best performance. For this purpose, various problem sizes are generated at random and then a robust calibration is applied to the parameters using the Taguchi method. In addition, two problems with different sizes are solved to evaluate the performance of the presented algorithm and to compare that performance with LINGO and also with the solution presented in previous work.     

On the location selection problem using analytic hierarchy process and multi-choice goal programming

Location selection is a crucial decision in cost/benefit analysis of restaurants, coffee shops and others. However, it is difficult to be solved because there are many conflicting multiple goals in the problem of location selection. In order to solve the problem, this study integrates analytic hierarchy process (AHP) and multi-choice goal programming (MCGP) as a decision aid to obtain an appropriate house from many alternative locations that better suit the preferences of renters under their needs. This study obtains weights from AHP and implements it upon each goal using MCGP for the location selection problem. According to the function of multi-aspiration provided by MCGP, decision makers can set multi-aspiration for each location goal to rank the candidate locations. Compared to the unaided selection processes, the integrated approach of AHP and MCGP is a better scientific and efficient method than traditional methods in finding a suitable location for buying or renting a house for business, especially under multiple qualitative and quantitative criteria within a shorter evaluation time. In addition, a real case is provided to demonstrate the usefulness of the proposed method. The results show that the proposed method is able to provide better quality decision than normal manual methods.     

Solving facilities location problem in the presence of alternative processing routes using a genetic algorithm

Facilities location problem deals with the optimization of location of manufacturing facilities like machines, departments, etc. in the shop floor. This problem greatly affects performance of a manufacturing system. It is assumed in this paper that there are multiple products to be produced on several machines. Alternative processing routes are considered for each product and the problem is to determine the processing route of each product and the location of each machine to minimize the total distance traveled by the materials within the shop floor. This paper presents a mixed-integer non-linear mathematical programming formulation to find optimal solution of this problem. A technique is used to linearize the formulated non-linear model. However, due to the NP-hardness of this problem, even the linearized model cannot be optimally solved by the conventional mathematical programming methods in a reasonable time. Therefore, a genetic algorithm is proposed to solve the linearized model. The effectiveness of the GA approach is evaluated with numerical examples. The results show that the proposed GA is both effective and efficient in solving the attempted problem.     

基于位置聚类和张量分解的Web服务推荐算法

基于服务质量（QoS）的Web服务推荐能在众多功能相似的Web服务中发现最能满足用户非功能需求的Web服务,但QoS属性值预测算法仍存在预测准确度不高和数据稀疏性的问题。针对以上问题,提出了一种基于位置聚类和分层张量分解的QoS预测算法ClustTD,该算法基于用户和服务的位置属性将用户和服务聚类成多个局部组,分别对局部组和全局的用户、服务和时间上下文进行张量建模和分解,将局部和全局张量分解的QoS预测值进行加权组合,同时考虑了局部和全局因素,获得最终QoS预测值。实验结果表明,该算法具有较高的QoS预测准确率和Web服务推荐质量,并能在一定程度上解决数据稀疏性问题。     

Hybrid graph convolutional networks with multi-head attention for location recommendation

World Wide Web - Recommending yet-unvisited points of interest (POIs) which may be of interest to users is one of the fundamental applications in location-based social networks. It mainly replies...     

Identifying representative users in matrix factorization-based recommender systems: application to solving the content-less new item cold-start problem

Matrix factorization has proven to be one of the most accurate recommendation approaches. However, it faces one major shortcoming: the latent features that result from the factorization are not directly interpretable. Providing interpretation for these features is important not only to help explain the recommendations presented to users, but also to understand the underlying relations between the users and the items. This paper consists of 2 contributions. First, we propose to automatically interpret features as users, referred to as representative users. This interpretation relies on the study of the matrices that result from the factorization and on their link with the original rating matrix. Such an interpretation is not only performed automatically, as it does not require any human expertise, but it also helps to explain the recommendations. The second proposition of this paper is to exploit this interpretation to alleviate the content-less new item cold-start problem. The experiments conducted on several benchmark datasets confirm that the features discovered by a Non-Negative Matrix Factorization can be interpreted as users and that representative users are a reliable source of information that allows to accurately estimate ratings on new items. They are thus a promising way to solve the new item cold-start problem.     

Improved heuristics for the regenerator location problem

Telecommunication systems use optical signals to transmit information. The strength of a signal in an optical network deteriorates and loses power as it goes farther from the source, mainly due to attenuation. Therefore, to enable the signal to arrive its intended destination with good quality, it is necessary to regenerate the signal periodically using regenerators. These components are relatively expensive and therefore it is desirable to deploy as few of them as possible in the network. In the regenerator location problem (RLP), we are given an undirected graph, positive edge lengths, and a parameter specifying the maximum length that a signal can travel before its quality deteriorates and regeneration is required. The problem consists in determining paths that connect all pairs of nodes in the graph and, if necessary, locating single regenerators in some of those nodes such that the signal never travels more than the maximum allowed distance without traversing a regenerator node. In this paper, we present new implementations of previous heuristics and two new heuristics—a GRASP and a biased random‐key genetic algorithm—for the RLP. Computational experiments comparing the proposed solution procedures with previous heuristics described in the literature illustrate the efficiency and effectiveness of our methods.