Gaussian variable neighborhood search for continuous optimization

Variable Neighborhood Search (VNS) has shown to be a powerful tool for solving both discrete and box-constrained continuous optimization problems. In this note we extend the methodology by allowing also to address unconstrained continuous optimization problems.Instead of perturbing the incumbent solution by randomly generating a trial point in a ball of a given metric, we propose to perturb the incumbent solution by adding some noise, following a Gaussian distribution. This way of generating new trial points allows one to give, in a simple and intuitive way, preference to some directions in the search space, or, contrarily, to treat uniformly all directions. Computational results show some advantages of this new approach.     

Intelligent video target tracking using an evolutionary particle filter based upon improved cuckoo search

The aim of this paper is to propose an evolutionary particle filter based upon improved cuckoo search algorithm which will overcome the sample impoverishment problem of generic particle filter. In our proposed method, improved cuckoo search (ICS) algorithm is embedded into particle filter (PF) framework. Improved cuckoo search algorithm uses levy flight for generating new particles in the solution and introduced randomness in samples by abandoning a fraction of these particles. The second important contribution in this article is introduction of new way for tackling scaling and rotational error in object tracking. Performance of proposed improved cuckoo particle filter is investigated and evaluated on synthetic and standard video sequences and compared with the generic particle filter and particle swarm optimization based particle filter. We show that object tracking using improved cuckoo particle filter provides more reliable and efficient tracking results than generic particle filter and PSO-particle filter. The proposed technique works for real time video objects tracking.     

邻域迭代重采样粒子滤波的纯方位目标跟踪

为了解决粒子滤波的非线性全局优化问题,基于重采样的思想是移除权重小的粒子,增加权重大的粒子数量,提出利用邻域搜索重采样的粒子滤波（NIRPF）进行目标跟踪。首先,预测粒子,并利用重要序列采样（SIS）给粒子赋权值;然后,在搜索后验概率密度的高概率区过程,更新单个粒子位置,利用高斯-邻域搜索迭代地加权所有粒子;最后,进行当前状态的估计。纯方位目标跟踪问题涉及两个静态观察器和非机动和机动两类目标。蒙特卡罗仿真结果验证了提出方法的有效性,与均方根容积卡尔曼滤波、容积粒子滤波和随机搜索的粒子滤波相比,提出的方法拥有更快的初始收敛速度,非机动目标和机动目标的根均方误差（RMSE）和时间根均方差（RTAMS）的评估更优。     

基于粒子滤波的模型自适应机动目标跟踪算法

针对当前机动目标跟踪领域中多模型算法存在的问题,提出一种基于粒子滤波的模型自适应机动目标跟踪算法.首先,依据前一时刻每个粒子采用的模型状态和模型间的状态转移概率,实现对当前时刻模型的采样;然后,将采样模型用于对当前粒子的预测,并根据当前时刻得到的量测数据实现对预测粒子权值的度量;最后,通过重采样策略和概率最大化原则完成对模型的合理选择和状态的有效估计.仿真实验验证了该算法的有效性.     

基于边缘粒子滤波的目标跟踪算法研究

为了提高目标跟踪过程中粒子滤波结果的精度,将边缘粒子滤波算法应用于目标跟踪。首先将目标运动状态向量划分为线性和非线性两个子向量,然后,采用卡尔曼滤波方法处理线性状态子向量,采用粒子滤波方法处理非线性状态子向量。使用边缘粒子滤波算法和标准粒子滤波算法对目标进行跟踪仿真。仿真结果表明：将边缘粒子滤波算法应用在目标跟踪过程中,能够取得更高的跟踪精度;时间复杂度增加仅6%;在粒子数相对较少的条件下,仍能够保持较好的滤波性能。     

A general variable neighborhood search for the swap-body vehicle routing problem

The Swap-Body Vehicle Routing Problem, a generalization of the well known Vehicle Routing Problem, can be stated as follows: the vehicle fleet consisting of trucks, semi-trailers, and swap bodies, is available at a single depot to serve a given set of customers. To serve a subset of customers, one may use either a truck carrying one swap body or a train (a truck with a semi-trailer attached to it) carrying two swap bodies. In both cases, a vehicle (a truck or a train) must perform a route starting and ending at the depot, so to satisfy demands of visited customers, maximal allowed route duration, allowed load on the used vehicle, and accessibility constraint of each customer. The accessibility constraint indicates whether a customer is allowed to be visited by a train or not. In addition, a set of swap locations is given where semi-trailers and swap bodies may be parked or swapped. The goal of the Swap-Body Vehicle Routing Problem is to minimize the total costs consisting of the fixed costs for using vehicles and costs for performing routes. In this paper, we propose two general variable neighborhood search heuristics to solve this problem. The quality of the proposed methods is evaluated on the instances provided by the organizers of VeRolog Solver Challenge 2014.     

A two phase variable neighborhood search for the multi-product inventory routing problem

We address a multi-product inventory routing problem and propose a two-phase Variable Neighborhood Search (VNS) metaheuristic to solve it. In the first phase, VNS is used to solve a capacitated vehicle routing problem at each period to find an initial solution without taking into account the inventory. In the second phase, we iteratively improve the initial solution while minimizing both the transportation and inventory costs. For this, we propose two different algorithms, a Variable Neighborhood Descent and a Variable Neighborhood Search. We present an heuristic and a Linear Programming formulation, which are applied after each local search move, to determine the amount of products to collect from each supplier at each period. During the exploration, we use priority rules for suppliers and vehicles, based on the current delivery schedule over the planning horizon. Computational results show the efficiency of the proposed two-phase approach.     

Ant colony optimization for bearings-only maneuvering target tracking in sensors network

In this paper, the problem of bearings-only maneuvering target tracking in sensors network is investigated. Two objectives are proposed and optimized by the ant colony optimization (ACO), then two kinds of node searching strategies of the ACO algorithm are presented. On the basis of the nodes determined by the ACO algorithm, the interacting multiple models extended Kalman filter (IMMEKF) for the multi-sensor bearings-only maneuvering target tracking is introduced. Simulation results indicate that the proposed ACO algorithm performs better than the Closest Nodes method. Furthermore, the Strategy 2 of the two given strategies is preferred in terms of the requirement of real time.     

Ant colony optimization for bearings-only maneuvering target tracking in sensors network

In this paper, the problem of bearings-only maneuvering target tracking in sensors network is investigated. Two objectives are proposed and optimized by the ant colony optimization （ACO）, then two kinds of node searching strategies of the ACO algorithm are presented. On the basis of the nodes determined by the ACO algorithm, the interacting multiple models extended Kalman filter （IMMEKF） for the multi-sensor bearings-only maneuvering target tracking is introduced. Simulation results indicate that the proposed ACO algorithm performs better than the Closest Nodes method. Furthermore, the Strategy 2 of the two given strategies is preferred in terms of the requirement of real time.     

基于特征约束和均值漂移的机动目标粒子跟踪

粒子滤波是适用于非线性非高斯系统下目标跟踪的强有力工具．ＭｉｒｏＳｏｔ足球机器人系统可以作为研究机动目标跟踪问题的平台．对此,在分析ＭｉｒｏＳｏｔ系统目标特征的基础上,提出一种基于目标特征约束的均值漂移粒子滤波算法,利用约束和优化的思想提高粒子的质量并减少其数量．对比实验表明,该方法有效地克服了传统粒子滤波的计算量和粒子退化问题,保证了多机动目标跟踪的准确性和实时性．     

Multi-dimensional visual tracking using scatter search particle filter

Multi-dimensional visual tracking (MVT) problems include visual tracking tasks where the system state is defined by a high number of variables corresponding to multiple model components and/or multiple targets. A MVT problem can be modeled as a dynamic optimization problem. In this context, we propose an algorithm which hybridizes particle filters (PF) and the scatter search (SS) metaheuristic, called scatter search particle filter (SSPF), where the optimization strategies from SS are embedded into the PF framework. Scatter search is a population-based metaheuristic successfully applied to several complex combinatorial optimization problems. The most representative optimization strategies from SS are both solution combination and solution improvement. Combination stage enables the solutions to share information about the problem to produce better solutions. Improvement stage makes also possible to obtain better solutions by exploring the neighborhood of a given solution. In this paper, we have described and evaluated the performance of the scatter search particle filter (SSPF) in MVT problems. Specifically, we have compared the performance of several state-of-the-art PF-based algorithms with SSPF algorithm in different instances of 2D articulated object tracking problem and 2D multiple object tracking. Some of these instances are from the CVBase’06 standard database. Experimental results show an important performance gain and better tracking accuracy in favour of our approach.     

Skewed general variable neighborhood search for the location routing scheduling problem

The integrated location routing scheduling problem is a variant of the well-known location routing problem. The location routing problem consists in selecting a set of depots to open and in building a set of routes from these depots, to serve a set of customers at minimum cost. In this variant, a vehicle can perform more than a single route in the planning period. As a consequence, the routes have to be scheduled within the workdays of each vehicle. The problem arises typically when routes are constrained to have a short duration. It happens for example within the boundaries of small geographic areas or in the transportation of perishable goods. In this paper, we propose a skewed general variable neighborhood search based heuristic to solve it. The algorithm is tested extensively and we show that it is efficient and provides the proven optimal solution in a significant number of cases. Moreover, it clearly outperforms a multi-start VND based heuristic that uses the same neighborhood structures.     

一种改进的基于Camshift的粒子滤波实时目标跟踪算法

为了能够快速和准确地跟踪运动目标,提出了一种改进的基于Camshift的粒子滤波算法。在粒子滤波框架下,首先对传统目标模型进行改进,提出一种新的融合目标颜色信息和运动信息的模型,以增强目标跟踪的稳健性和准确性;同时为了提高跟踪的效率,将一种改进的Camshift算法嵌入到粒子滤波中,用来重新分配随机粒子样本,使之向目标状态的最大后验概率密度方向移动。实验结果表明,与传统的粒子滤波算法或Camshift算法相比,该方法能有效处理目标快速运动或背景存在强干扰等情况,实现对目标快速和稳健的跟踪。     

A population-based variable neighborhood search for the single machine total weighted tardiness problem

This paper investigates the single machine total weighted tardiness problem, in which a set of independent jobs with distinct processing times, weights, and due dates are to be scheduled on a single machine to minimize the sum of weighted tardiness of all jobs. This problem is known to be strongly NP-hard, and thus provides a challenging area for metaheuristics. A population-based variable neighborhood search (PVNS) algorithm is developed to solve it. This algorithm differs from the basic variable neighborhood search (VNS). First, the PVNS consists of a number of iterations of the basic VNS, and in each iteration a population of solutions is used to simultaneously generate multiple trial solutions in a neighborhood so as to improve the search diversification. Second, the PVNS adopts a combination of path-relinking, variable depth search and tabu search to act as the local search procedure so as to improve the search intensification. Computational experiments show that the proposed PVNS algorithm can obtain the optimal or best known solutions within a reasonable computation time for all standard benchmark problem instances from the literature.     

A hybrid occlusion free object tracking method using particle filter and modified galaxy based search meta-heuristic algorithm

An important problem in tracking methods is how to manage the changes in object appearance, such as illumination changes, partial/full occlusion, scale, and pose variation during the tracking process. In this paper, we propose an occlusion free object tracking method together with a simple adaptive appearance model. The proposed appearance model which is updated at the end of each time step includes three components: the first component consists of a fixed template of target object, the second component shows rapid changes in object appearance, and the third one maintains slow changes generated along the object path. The proposed tracking method not only can detect occlusion and handle it, but also it is robust against changes in the object appearance model. It is based on particle filter which is a robust technique in tracking and handles non-linear and non-Gaussian problems. We have also employed a meta-heuristic approach that is called Modified Galaxy based Search Algorithm (MGbSA), to reinforce finding the optimum state in the particle filter state space. The proposed method was applied to some benchmark videos and its results were satisfactory and better than results of related works.     

自适应变邻域混沌搜索微粒群算法

针对局部地形复杂、振荡强烈的函数优化精度难以提高的问题,提出一种自动调整邻域搜索范围和方向的自适应变邻域混沌搜索微粒群算法（AVNC－PSO）。优化初期首先由基本PSO算法进行粗调,当种群收敛于局部最优时,选择飞行停滞且聚集程度高的粒子向不同方向的邻域内进行混沌搜索,搜索方向和粒子偏移量根据粒子与收敛中心的距离和混沌变量的值共同确定。数值仿真表明,该算法能够使局部搜索更精确,有效改善基本PSO算法优化精度不高的弱点。     

A variable neighborhood search algorithm for the surgery tactical planning problem

We address the tactical planning problem of surgeries that consists in building an admission plan of patients over a medium-term horizon planning so as to minimize over and under utilization of several resources such as operating theaters, beds and nursing care, compared with their target level of utilization. The problem is formulated as a mixed integer linear program for which exact solution methods fail to find an optimal solution in a reasonable execution time. We develop a Variable Neighborhood Search algorithm and show its ability to provide high quality solutions in short computational running times compared with CPLEX for numerous real-sized instances based on the surgery planning problem in a Dutch cardiothoracic center. Furthermore, with few parameters' settings and low computational memory requirements, this approach may easily be implemented in a decision support system for hospitals.     

Variable neighborhood search for the pharmacy duty scheduling problem

In this paper, we study on the Pharmacy Duty Scheduling (PDS) problem, where a subset of pharmacies should be on duty on national holidays, at weekends and at nights in order to be able to satisfy the emergency drug needs of the society. PDS problem is a multi-period p-median problem with special side constraints and it is an NP-Hard problem. We propose four Variable Neighborhood Search (VNS) heuristics. The first one is the basic version, BVNS. The latter two, Variable Neighborhood Decomposition Search (VNDS) and Variable Neighborhood Restricted Search (VNRS), aim to obtain better results in less computing time by decomposing or restricting the search space. The last one, Reduced VNS (RVNS), is for obtaining good initial solutions rapidly for BVNS, VNDS and VNRS. We test BVNS, VNRS and VNDS heuristics on randomly generated instances and report the computational test results. We also use VNS heuristics on real data for the pharmacies in central İzmir and obtain significant improvements.     

The double traveling salesman problem with multiple stacks: A variable neighborhood search approach

The double traveling salesman problem with multiple stacks (DTSPMS) is a vehicle routing problem that consists on finding the minimum total length tours in two separated networks, one for pickups and one for deliveries. A set of orders is given, each one consisting of a pickup location and a delivery location, and it is required to send an item from the former location to the latter one. Repacking is not allowed, but collected items can be packed in several rows in such a way that each row must obey the LIFO principle. In this paper, a variable neighborhood search approach using four new neighborhood structures is presented to solve the problem.     

A hybrid variable neighborhood tabu search heuristic for the vehicle routing problem with multiple time windows

This paper presents a new hybrid variable neighborhood-tabu search heuristic for the Vehicle Routing Problem with Multiple Time windows. It also proposes a minimum backward time slack algorithm applicable to a multiple time windows environment. This algorithm records the minimum waiting time and the minimum delay during route generation and adjusts the arrival and departure times backward. The implementation of the proposed heuristic is compared to an ant colony heuristic on benchmark instances involving multiple time windows. Computational results on newly generated instances are provided.