Measuring potential gains from specialization under non-convex technologies

In this paper, the Free Coordination Hull (FCH) approach developed by Green and Cook (2004) is combined with the Free Disposal Hull (FDH) model to detect potential gains from specialization. As a non-convex approach that allows both directly observed and summed Decision Making Units to define the production technology, FCH is the relevant model for analysing optimal reallocation of activity among smaller and more specialized units. Indeed in more traditional Data Envelopment Analysis models the convexity assumption precludes the possibility of detecting potential gains from specialization and can only reveal economies of scope. Therefore non-convex technologies are required to model diseconomies of scope. On the basis of FDH and FCH technologies, an overall efficiency measure is decomposed into three components, namely: technical, size and specialization efficiencies. A 2003 database of French farms is used as an illustration. Results indicate that input inefficiency in the agricultural sector is mainly driven by a lack of specialization, which represents approximately 50% of the overall inefficiency.     

Allocative efficiency of technically inefficient production units

We discuss how to measure allocative efficiency without presuming technical efficiency. This is relevant when it is easier to introduce reallocations than improvements of technical efficiency. We compare the approach to the traditional one of assuming technical efficiency before measuring allocative efficiency. In particular, we develop necessary and sufficient conditions in the technology to ensure consistent measures, we suggest alternative interpretations of the approaches, and we relate them to motivational and organizational change perspectives.     

Efficiency measurement with a non-convex free disposal hull technology

We investigate the basic monotonicity properties of least-distance (in)efficiency measures on the class of non-convex FDH (free disposable hull) technologies. We show that any known FDH least-distance measure violates strong monotonicity over the strongly (Pareto-Koopmans) efficient frontier. Taking this result into account, we develop a new class of FDH least-distance measures that satisfy strong monotonicity and show that the developed (in)efficiency measurement framework has a natural profit interpretation.     

Trajectory planning based on non-convex global optimization for serial manipulators

To perform specific tasks in dynamic environments, robots are required to rapidly update trajectories according to changing factors. A continuous trajectory planning methodology for serial manipulators based on non-convex global optimization is presented in this paper. First, a kinematic trajectory planning model based on non-convex optimization is constructed to balance motion rapidity and safety. Then, a model transformation method for the non-convex optimization model is presented. In this way, the accurate global solution can be obtained with an iterative solver starting from arbitrary initializations, which can greatly improve the computational accuracy and efficiency. Furthermore, an efficient initialization method for the iterative solver based on multivariable-multiple regression is presented, which further speeds up the solution process. The results show that trajectory planning efficiency is significantly enhanced by model transformation and initialization improvement for the iterative solver. Consequently, real-time continuous trajectory planning for serial manipulators with many degrees of freedom can be achieved, which lays a basis for performing dynamic tasks in complex environments.     

Local and global returns to scale in performance measurement

In data envelopment analysis, the type of local returns to scale (RTS) exhibited by a technically efficient unit indicates whether an increase or reduction of the scale of operations could improve the productivity of the unit. One of the approaches to testing RTS is based on the comparison of the efficiency of the unit in specially constructed reference technologies. It has been suggested that this approach is equally suitable for convex and non-convex, including the free disposal hull, technologies. In this paper, we construct examples that show that this suggestion in the case of non-convex technologies is not correct. We show that the type of RTS obtained by this approach is not a local, but global, characteristic of the technology, as it indicates the direction to the most productive scale size of the unit. In non-convex technologies, the local and global classifications of RTS are generally different.     

A Globally Convergent and Efficient Method for Unconstrained Discrete-Time Optimal Control

Shift schemes are commonly used in non-convex situations when solving unconstrained discrete-time optimal control problems by the differential dynamic programming (DDP) method. However, the existing shift schemes are inefficient when the shift becomes too large. In this paper, a new method of combining the DDP method with a shift scheme and the steepest descent method is proposed to cope with non-convex situations. Under certain assumptions, the proposed method is globally convergent and has q-quadratic local conve rgence. Extensive numerical experiments on many test problems in the literature are reported. These numerical results illustrate the robustness and efficiency of the proposed method.     

Location of Two Facilities with Minimal Separation

In this technical note it is shown how a problem of location of two facilities may be formulated as a non-convex optimization problem, and an algorithm is provided to determine an ?-optimal solution.     

Projection Method Approach for General Regularized Non-convex Variational Inequalities

In this paper, we investigate or analyze non-convex variational inequalities and general non-convex variational inequalities. Two new classes of non-convex variational inequalities, named regularized non-convex variational inequalities and general regularized non-convex variational inequalities, are introduced, and the equivalence between these two classes of non-convex variational inequalities and the fixed point problems are established. A projection iterative method to approximate the solutions of general regularized non-convex variational inequalities is suggested. Meanwhile, the existence and uniqueness of solution for general regularized non-convex variational inequalities is proved, and the convergence analysis of the proposed iterative algorithm under certain conditions is studied.     

Robust solutions of quadratic optimization over single quadratic constraint under interval uncertainty

In this paper we examine non-convex quadratic optimization problems over a quadratic constraint under unknown but bounded interval perturbation of problem data in the constraint and develop criteria for characterizing robust (i.e. uncertainty-immunized) global solutions of classes of non-convex quadratic problems. Firstly, we derive robust solvability results for quadratic inequality systems under parameter uncertainty. Consequently, we obtain characterizations of robust solutions for uncertain homogeneous quadratic problems, including uncertain concave quadratic minimization problems and weighted least squares. Using homogenization, we also derive characterizations of robust solutions for non-homogeneous quadratic problems.     

Numerical approximation of young measuresin non-convex variational problems

Summary. In non-convex optimisation problems, in particular in non-convex variational problems, there usually does not exist any classical solution but only generalised solutions which involve Young measures. In this paper, first a suitable relaxation and approximation theory is developed together with optimality conditions, and then an adaptive scheme is proposed for the efficient numerical treatment. The Young measures solving the approximate problems are usually composed only from a few atoms. This is the main argument our effective active-set type algorithm is based on. The support of those atoms is estimated from the Weierstrass maximum principle which involves a Hamiltonian whose good guess is obtained by a multilevel technique. Numerical experiments are performed in a one-dimensional variational problem and support efficiency of the algorithm. Received November 26, 1997 / Published online September 24, 1999     

About the non-convex optimization problem induced by non-positive semidefinite kernel learning

During the last years, kernel based methods proved to be very successful for many real-world learning problems. One of the main reasons for this success is the efficiency on large data sets which is a result of the fact that kernel methods like support vector machines (SVM) are based on a convex optimization problem. Solving a new learning problem can now often be reduced to the choice of an appropriate kernel function and kernel parameters. However, it can be shown that even the most powerful kernel methods can still fail on quite simple data sets in cases where the inherent feature space induced by the used kernel function is not sufficient. In these cases, an explicit feature space transformation or detection of latent variables proved to be more successful. Since such an explicit feature construction is often not feasible for large data sets, the ultimate goal for efficient kernel learning would be the adaptive creation of new and appropriate kernel functions. It can, however, not be guaranteed that such a kernel function still leads to a convex optimization problem for Support Vector Machines. Therefore, we have to enhance the optimization core of the learning method itself before we can use it with arbitrary, i.e., non-positive semidefinite, kernel functions. This article motivates the usage of appropriate feature spaces and discusses the possible consequences leading to non-convex optimization problems. We will show that these new non-convex optimization SVM are at least as accurate as their quadratic programming counterparts on eight real-world benchmark data sets in terms of the generalization performance. They always outperform traditional approaches in terms of the original optimization problem. Additionally, the proposed algorithm is more generic than existing traditional solutions since it will also work for non-positive semidefinite or indefinite kernel functions.     

求解鲁棒主成分分析的新交替下降方向法

鲁棒主成分分析作为统计与数据科学领域的基本工具已被广泛研究,其核心原理是把观测数据分解成低秩部分和稀疏部分.本文基于鲁棒主成分分析的非凸模型,提出了一种新的基于梯度方法和非单调搜索技术的高斯型交替下降方向法.在新算法中,交替更新低秩部分和稀疏部分相关的变量,其中低秩部分的变量是利用一步带有精确步长的梯度下降法进行更新,...     

A branch-reduce-cut algorithm for the global optimization of probabilistically constrained linear programs

We consider probabilistically constrained linear programs with general distributions for the uncertain parameters. These problems involve non-convex feasible sets. We develop a branch-and-bound algorithm that searches for a global optimal solution to this problem by successively partitioning the non-convex feasible region and by using bounds on the objective function to fathom inferior partition elements. This basic algorithm is enhanced by domain reduction and cutting plane strategies to reduce the size of the partition elements and hence tighten bounds. The proposed branch-reduce-cut algorithm exploits the monotonicity properties inherent in the problem, and requires solving linear programming subproblems. We provide convergence proofs for the algorithm. Some illustrative numerical results involving problems with discrete distributions are presented.     

Statistical inference for DEA estimators of directional distances

In productivity and efficiency analysis, the technical efficiency of a production unit is measured through its distance to the efficient frontier of the production set. The most familiar non-parametric methods use Farrell–Debreu, Shephard, or hyperbolic radial measures. These approaches require that inputs and outputs be non-negative, which can be problematic when using financial data. Recently, Chambers et al. (1998) have introduced directional distance functions which can be viewed as additive (rather than multiplicative) measures efficiency. Directional distance functions are not restricted to non-negative input and output quantities; in addition, the traditional input and output-oriented measures are nested as special cases of directional distance functions. Consequently, directional distances provide greater flexibility. However, until now, only free disposal hull (FDH) estimators of directional distances (and their conditional and robust extensions) have known statistical properties (Simar and Vanhems, 2012). This paper develops the statistical properties of directional d estimators, which are especially useful when the production set is assumed convex. We first establish that the directional Data Envelopment Analysis (DEA) estimators share the known properties of the traditional radial DEA estimators. We then use these properties to develop consistent bootstrap procedures for statistical inference about directional distance, estimation of confidence intervals, and bias correction. The methods are illustrated in some empirical examples.     

Stochastic relaxed inertial forward-backward-forward splitting for monotone inclusions in Hilbert spaces

Computational Optimization and Applications - We develop a globalized Proximal Newton method for composite and possibly non-convex minimization problems in Hilbert spaces. Additionally, we impose...     

Investigating a hybrid simulated annealing and local search algorithm for constrained optimization

Constrained Optimization Problems (COP) often take place in many practical applications such as kinematics, chemical process optimization, power systems and so on. These problems are challenging in terms of identifying feasible solutions when constraints are non-linear and non-convex. Therefore, finding the location of the global optimum in the non-convex COP is more difficult as compared to non-convex bound-constrained global optimization problems. This paper proposes a Hybrid Simulated Annealing method (HSA), for solving the general COP. HSA has features that address both feasibility and optimality issues and here, it is supported by a local search procedure, Feasible Sequential Quadratic Programming (FSQP). We develop two versions of HSA. The first version (HSAP) incorporates penalty methods for constraint handling and the second one (HSAD) eliminates the need for imposing penalties in the objective function by tracing feasible and infeasible solution sequences independently. Numerical experiments show that the second version is more reliable in the worst case performance.     

A completely positive representation of 0–1 linear programs with joint probabilistic constraints

In this paper, we study 0–1 linear programs with joint probabilistic constraints. The constraint matrix vector rows are assumed to be independent, and the coefficients to be normally distributed. Our main results show that this non-convex problem can be approximated by a convex completely positive problem. Moreover, we show that the optimal values of the latter converge to the optimal values of the original problem. Examples randomly generated highlight the efficiency of our approach.     

Hybrid algorithm for a vendor managed inventory system in a two-echelon supply chain

In this paper we address the issue of vendor managed inventory (VMI) by considering a two-echelon single vendor/multiple buyer supply chain network. We try to find the optimal sales quantity by maximizing profit, given as a nonlinear and non-convex objective function. For such complicated combinatorial optimization problems, exact algorithms and optimization commercial software such as LINGO are inefficient, especially on practical-size problems. In this paper we develop a hybrid genetic/simulated annealing algorithm to deal with this nonlinear problem. Our results demonstrate that the proposed hybrid algorithm outperforms previous methodologies and achieves more robust solutions.     

Unstructured mesh finite difference/finite element method for the 2D time-space Riesz fractional diffusion equation on irregular convex domains

Fractional differential equations are powerful tools to model the non-locality and spatial heterogeneity evident in many real-world problems. Although numerous numerical methods have been proposed, most of them are limited to regular domains and uniform meshes. For irregular convex domains, the treatment of the space fractional derivative becomes more challenging and the general methods are no longer feasible. In this work, we propose a novel numerical technique based on the Galerkin finite element method (FEM) with an unstructured mesh to deal with the space fractional derivative on arbitrarily shaped convex and non-convex domains, which is the most original and significant contribution of this paper. Moreover, we present a second order finite difference scheme for the temporal fractional derivative. In addition, the stability and convergence of the method are discussed and numerical examples on different irregular convex domains and non-convex domains illustrate the reliability of the method. We also extend the theory and develop a computational model for the case of a multiply-connected domain. Finally, to demonstrate the versatility and applicability of our method, we solve the coupled two-dimensional fractional Bloch–Torrey equation on a human brain-like domain and exhibit the effects of the time and space fractional indices on the behaviour of the transverse magnetization.     

Genetic Algorithm based technique for solving Chance Constrained Problems

Management and measurement of risk is an important issue in almost all areas that require decisions to be made under uncertain information. Chance Constrained Programming (CCP) have been used for modelling and analysis of risks in a number of application domains. However, the resulting mathematical problems are non-trivial to represent using algebraic modelling languages and pose significant computational challenges due to their non-linear, non-convex, and the stochastic nature. We develop and implement C++ classes to represent such CCP problems. We propose a framework consisting of Genetic Algorithm and Monte Carlo Simulation in order to process the problems. The non-linear and non-convex nature of the CCP problems are processed using Genetic Algorithm, whereas the stochastic nature is addressed through Simulation. The computational investigations have shown that the framework can efficiently represent and obtain good solutions for seven test problems.