A new reverse reduce-error ensemble pruning algorithm

Although greedy algorithms possess high efficiency, they often receive suboptimal solutions of the ensemble pruning problem, since their exploration areas are limited in large extent. And another marked defect of almost all the currently existing ensemble pruning algorithms, including greedy ones, consists in: they simply abandon all of the classifiers which fail in the competition of ensemble selection, causing a considerable waste of useful resources and information. Inspired by these observations, an interesting greedy Reverse Reduce-Error (RRE) pruning algorithm incorporated with the operation of subtraction is proposed in this work. The RRE algorithm makes the best of the defeated candidate networks in a way that, the Worst Single Model (WSM) is chosen, and then, its votes are subtracted from the votes made by those selected components within the pruned ensemble. The reason is because, for most cases, the WSM might make mistakes in its estimation for the test samples. And, different from the classical RE, the near-optimal solution is produced based on the pruned error of all the available sequential subensembles. Besides, the backfitting step of RE algorithm is replaced with the selection step of a WSM in RRE. Moreover, the problem of ties might be solved more naturally with RRE. Finally, soft voting approach is employed in the testing to RRE algorithm. The performances of RE and RRE algorithms, and two baseline methods, i.e., the method which selects the Best Single Model (BSM) in the initial ensemble, and the method which retains all member networks of the initial ensemble (ALL), are evaluated on seven benchmark classification tasks under different initial ensemble setups. The results of the empirical investigation show the superiority of RRE over the other three ensemble pruning algorithms.     

Elitist clonal selection algorithm for optimal choice of free knots in B-spline data fitting

Data fitting with B-splines is a challenging problem in reverse engineering for CAD/CAM, virtual reality, data visualization, and many other fields. It is well-known that the fitting improves greatly if knots are considered as free variables. This leads, however, to a very difficult multimodal and multivariate continuous nonlinear optimization problem, the so-called knot adjustment problem. In this context, the present paper introduces an adapted elitist clonal selection algorithm for automatic knot adjustment of B-spline curves. Given a set of noisy data points, our method determines the number and location of knots automatically in order to obtain an extremely accurate fitting of data. In addition, our method minimizes the number of parameters required for this task. Our approach performs very well and in a fully automatic way even for the cases of underlying functions requiring identical multiple knots, such as functions with discontinuities and cusps. To evaluate its performance, it has been applied to three challenging test functions, and results have been compared with those from other alternative methods based on AIS and genetic algorithms. Our experimental results show that our proposal outperforms previous approaches in terms of accuracy and flexibility. Some other issues such as the parameter tuning, the complexity of the algorithm, and the CPU runtime are also discussed.     

An adaptive local search algorithm for vehicle routing problem with simultaneous and mixed pickups and deliveries

The Vehicle Routing Problem with Simultaneous Pickup and Delivery (VRPSPD) is an extension to the classical Vehicle Routing Problem (VRP), where customers may both receive and send goods simultaneously. The Vehicle Routing Problem with Mixed Pickup and Delivery (VRPMPD) differs from the VRPSPD in that the customers may have either pickup or delivery demand. However, the solution approaches proposed for the VRPSPD can be directly applied to the VRPMPD. In this study, an adaptive local search solution approach is developed for both the VRPSPD and the VRPMPD, which hybridizes a Simulated Annealing inspired algorithm with Variable Neighborhood Descent. The algorithm uses an adaptive threshold function that makes the algorithm self-tuning. The proposed approach is tested on well-known VRPSPD and VRPMPD benchmark instances derived from the literature. The computational results indicate that the proposed algorithm is effective in solving the problems in reasonable computation time.     

Understanding value creation in closed loop supply chains – Past findings and future directions

The topic of value creation through the recovery of returned products in closed loop supply chains is scattered across various bodies of literature. We undertake a systematic literature review of 144 articles in relevant green, reverse and closed loop supply chain literature to synchronize existing knowledge on value creation. Value manifestations of four types of value, namely economic, environmental, information and customer value, are identified. Value adding concepts from the forward- and reverse supply chain may leverage the process of value creation. They are classified into six subclasses, namely partnerships and collaboration, product design characteristics, service concepts, IT solutions, supply chain processes and organizational characteristics. We present a conceptual framework on a strategic level. In this way companies can create competitive advantages by closing the loop. The results of the literature analysis suggest avenues for future research on the operational and strategic level.     

基于TMS320VC5410的最小系统设计

TMS320VC5410是TI公司生产的一款定点数字信号处理(DSP)芯片。基于DSP的最小系统是以其简洁的硬件结构来实现DSP的一般功能,广泛应用于工业生产、教学、科研、电子设计等领域。主要介绍了基于TMS320VC5410的最小系统硬件电路设计方案,阐述了重要电路部分的抗干扰设计和软件调试过程中应注意的若干问题。最后提出了简单易行的软件调试方法以及部分程序代码。     

Influence of Reaction Parameters on the Hydrogenolysis of Hydroxymatairesinol Over Carbon Nanofibre Supported Palladium Catalysts

The density functional theory (DFT/B3LYP) calculations were applied to investigate the interaction of a Pt₆ particle with the ZSM-5 zeolite framework. The electronic structure of the metal particle is strongly affected by the interaction with basic framework oxygens and acid sites of the zeolite support. Adsorption on basic sites (E_ads = 6 kcal/mol) favors the formation of the electron enriched metal cluster. Interaction of the platinum cluster with the acid site characterized by stabilization energy of 47 kcal/mol results in oxidation of the metal particle and suppression of Brønsted acidity of the support. The hypothesis is put forward that the oxidized platinum particle can function as an active site for the alkane isomerisation on platinum supported high silica zeolites.     

Disassembly sequence planning in a disassembly cell context

In this paper a two-phase approach is proposed for determining the optimal disassembly sequence when the disassembly system has a cellular configuration. Operations are first grouped into cells based on the resources they require with the goal of minimizing machine acquisition costs. The aim is to group together those operations that use similar equipment in order to achieve good utilization levels of such equipment. A maximum cell size may be imposed. Once the cells have been formed and the operations have been assigned to them, a metaheuristic algorithm (namely GRASP) is used to search for the disassembly sequence for each product that leads to the minimum number of intercellular movements. To account for uncertainty regarding the condition in which the product may arrive, each disassembly task is assumed to be required with a certain probability, regardless of the other tasks. AND/OR precedence relations among the disassembly tasks are also considered. The proposed approach is illustrated on a randomly generated disassembly problem.     

Targeting optimum resource allocation using reverse problem formulations and property clustering techniques

Decoupling the constitutive equations from the balance and constraint equations allows for reformulating a conventional forward problem into two reverse problems. The first reverse problem is the reverse of a simulation problem, where the process model is solved in terms of the constitutive (synthesis/design) variables instead of the process variables, thus providing the synthesis/design targets. The second reverse problem (reverse property prediction) solves the constitutive equations to identify unit operations, operating conditions and/or products by matching the synthesis/design targets. Visualization of the problem is achieved by employing recently developed property clustering techniques, which allows a high-dimensional problem to be visualized in two or three dimensions. The clusters by definition satisfy intra-stream and inter-stream conservation through linear “mixing” rules, which allows for the development of consistent additive rules along with their ternary representation.     

Debugging programs that use macro-oriented data abstractions

Enhanced C (EC) is a set-oriented, extensible, C-like language. EC uses data abstractions to define new types. These data abstractions, called clusters, are macro-like devices that perform substitution on the typed syntax tree. Debugging programs that use clusters raise problems that are not encountered in ordinary programming languages. At compile time there is a need to determine and report whether the macro expansion will result in a legal program before this expansion actually takes place. At run-time the problems are how to account for replaced statements and how to handle variables whose types have been established by the clusters, variables that disappear, or variables whose names have been changed. This article presents these problems and their solutions as implemented by the EC compiler and the EC symbolic debugger. Similar debugging problems appear in other languages: The need to handle variables at run time is common to all languages that support data abstraction even if the abstractions are procedure oriented; also, a mild form of the problem of the replaced statement appears in inline procedure substitution of Ada. The solutions developed for the EC debugger apply to these cases as well.     

Algorithm for finding all k nearest neighbors

A very simple algorithm for computing all k nearest neighbors in 2-D is presented. The method does not rely on complicated forms of tessellation, it only requires simple data binning for fast range searching. Its applications range from scattered data interpolation to reverse engineering.