Monte Carlo planning for active object classification

Classifying objects in complex unknown environments is a challenging problem in robotics and is fundamental in many applications. Modern sensors and sophisticated perception algorithms extract rich 3D textured information, but are limited to the data that are collected from a given location or path. We are interested in closing the loop around perception and planning, in particular to plan paths for better perceptual data, and focus on the problem of planning scanning sequences to improve object classification from range data. We formulate a novel time-constrained active classification problem and propose solution algorithms that employ a variation of Monte Carlo tree search to plan non-myopically. Our algorithms use a particle filter combined with Gaussian process regression to estimate joint distributions of object class and pose. This estimator is used in planning to generate a probabilistic belief about the state of objects in a scene, and also to generate beliefs for predicted sensor observations from future viewpoints. These predictions consider occlusions arising from predicted object positions and shapes. We evaluate our algorithms in simulation, in comparison to passive and greedy strategies. We also describe similar experiments where the algorithms are implemented online, using a mobile ground robot in a farm environment. Results indicate that our non-myopic approach outperforms both passive and myopic strategies, and clearly show the benefit of active perception for outdoor object classification.     

Scheduling algorithms for multicast traffic in TDM/WDM networks with arbitrary tuning latencies

We consider all-optical Time Division Multiplexing (TDM)/Wavelength Division Multiplexing (WDM) broadcast and select networks with slotted operation. Each network access node is equipped with one fixed transmitter and one tunable receiver; tuning times are not negligible with respect to the fixed size slot time. We discuss efficient scheduling algorithms to assign TDM/WDM slots to multicast traffic in such networks. The problem is shown to be NP-hard; thus, heuristic algorithms based on the Tabu Search meta-heuristic are proposed, and their performance are assessed using randomly created request matrices based on two types of multicast traffic patterns. We show that significant advantages can be obtained by using these novel algorithms with respect to simpler greedy algorithms, even when restricting CPU times to realistic values to make the algorithms of practical use.     

Incremental Beam search

Beam search is a heuristic search algorithm that explores a state-space graph by expanding w most promising nodes at each level (depth) of the graph, where w is called the beam-width which is taken as input from the user. The quality of the solution produced by beam search does not always monotonically improve with the increase in beam-width making it difficult to choose an appropriate beam-width for effective use. We present an algorithm called Incremental Beam Search (IncB) which guarantees monotonicity, and is also anytime in nature. Experimental results on the sliding-tile puzzle, the traveling salesman, and the single-machine scheduling problems show that IncB significantly outperforms basic monotonic methods such as iterative widening beam search as well as some of the state-of-the-art anytime heuristic search algorithms in terms of the quality of the solution produced at the end as well as the anytime performance.     

Constructive heuristic algorithms for the open shop problem

In this paper we consider constructive heuristic algorithms for the open shop problem with minimization of the schedule length. By means of investigations of the structure of a feasible solution two types of heuristic algorithms are developed: construction of a rank-minimal schedule by solving successively weighted maximum cardinality matching problems and construction of an approximate schedule by applying insertion techniques combined with beam search. All presented algorithms are tested on benchmark problems from the literature. Our computational results demonstrate the excellent solution quality of our insertion algorithm, especially for greater job and machine numbers. For 29 of 30 benchmark problems with at least 10 jobs and 10 machines we improve the best known values obtained by tabu search.     

Using genetic algorithms to generate test sequences for complex timed systems

The generation of test data for state-based specifications is a computationally expensive process. This problem is magnified if we consider that time constraints have to be taken into account to govern the transitions of the studied system. The main goal of this paper is to introduce a complete methodology, supported by tools, that addresses this issue by representing the test data generation problem as an optimization problem. We use heuristics to generate test cases. In order to assess the suitability of our approach we consider two different case studies: a communication protocol and the scientific application BIPS3D. We give details concerning how the test case generation problem can be presented as a search problem and automated. Genetic algorithms (GAs) and random search are used to generate test data and evaluate the approach. GAs outperform random search and seem to scale well as the problem size increases. It is worth to mention that we use a very simple fitness function that can be easily adapted to be used with other evolutionary search techniques.     

Optimal discrete nonlinear filters of the objects’s order and their Gaussian approximations

We consider the problem of devising effici ent algorithms of estimating the current state of nonlinear stochastic dynamical systems by observing their outputs. We analyze the shortcomings of existing discrete Markov filtering approaches. For a previously suggested nonlinear filter of optimal structure and a new two-step filter, whose orders equal the number of components of the estimated state vector, we give procedures for both exact and approximately-analytic computation of their structural functions. We consider a way to enhance the precision of the designed filters with suboptimal structure by optimizing their auxiliary parameters once again. We compare the suggested and known estimation algorithms.     

Searching the Web with Queries

In this paper we study the problem of searching the Web with online learning algorithms. We consider that Web documents can be represented by vectors of n boolean attributes. A search engine is viewed as a learner, and a user is viewed as a teacher. We investigate the number of queries a search engine needs from the user to search for a collection of Web documents. We design several efficient learning algorithms to search for any collection of documents represented by a disjunction (or a conjunction) of relevant attributes with the help of membership queries or equivalence queries.     

Heuristics for search sequencing under time-dependent probabilities of existence

We consider the problem of determining a route of a search resource to search visually multiple areas in which targets are expected to be located. It is assumed that the probability a target exists in each area is given as a result of target detection operations and that the probability decreases as time passes. It is necessary to search the areas using a search resource, and identify the exact locations of the targets. We propose heuristic algorithms including a simulated annealing (SA) algorithm for the search sequencing problem. Since the search sequence must be determined as quickly as possible not to delay the search, heuristics for search sequencing should not take too much time. We introduce a new neighborhood generation method and a new parameter for an easier control of the overall computation time in the SA algorithm. A series of computational experiments is performed for evaluating the suggested algorithms, and results are reported.     

基于再生核希尔伯特空间映射的高维数据特征选择优化算法*

现有过滤型特征选择算法并未考虑非线性数据的内在结构,从而分类准确率远远低于封装型算法,对此提出一种基于再生核希尔伯特空间映射的高维数据特征选算法。首先,基于分支定界法建立搜索树,并对其进行搜索;然后,基于再生核希尔伯特空间映射分析非线性数据的内部结构;最终,根据数据集的内部结构选择最优的距离计算方法。对比仿真实验结果表明,本方法与封装型特征选择算法具有接近的分类准确率,同时在计算效率上具有明显的优势,适用于大数据分析。     

The impact of parametrization in memetic evolutionary algorithms

Memetic (evolutionary) algorithms integrate local search into the search process of evolutionary algorithms. As computational resources have to be spread adequately among local and evolutionary search, one has to care about when to apply local search and how much computational effort to devote to local search. Often local search is called with a fixed frequency and run for a fixed number of iterations, the local search depth. There is empirical evidence that these parameters have a significant impact on performance, but a theoretical understanding as well as concrete design guidelines are missing.     

A genetic algorithm approach for the single machine scheduling problem with linear earliness and quadratic tardiness penalties

In this paper, we consider the single machine scheduling problem with linear earliness and quadratic tardiness costs, and no machine idle time. We propose a genetic approach based on a random key alphabet. Several genetic algorithms based on this approach are presented. These versions differ on the generation of the initial population, as well as on the use of local search. The proposed procedures are compared with existing heuristics, as well as with optimal solutions for the smaller instance sizes.     

Block matching algorithm for motion estimation based on Artificial Bee Colony (ABC)

Block matching (BM) motion estimation plays a very important role in video coding. In a BM approach, image frames in a video sequence are divided into blocks. For each block in the current frame, the best matching block is identified inside a region of the previous frame, aiming to minimize the sum of absolute differences (SAD). Unfortunately, the SAD evaluation is computationally expensive and represents the most consuming operation in the BM process. Therefore, BM motion estimation can be approached as an optimization problem, where the goal is to find the best matching block within a search space. The simplest available BM method is the full search algorithm (FSA) which finds the most accurate motion vector through an exhaustive computation of SAD values for all elements of the search window. Recently, several fast BM algorithms have been proposed to reduce the number of SAD operations by calculating only a fixed subset of search locations at the price of poor accuracy. In this paper, a new algorithm based on Artificial Bee Colony (ABC) optimization is proposed to reduce the number of search locations in the BM process. In our algorithm, the computation of search locations is drastically reduced by considering a fitness calculation strategy which indicates when it is feasible to calculate or only estimate new search locations. Since the proposed algorithm does not consider any fixed search pattern or any other movement assumption as most of other BM approaches do, a high probability for finding the true minimum (accurate motion vector) is expected. Conducted simulations show that the proposed method achieves the best balance over other fast BM algorithms, in terms of both estimation accuracy and computational cost.     

Computational noise effects on adaptive filter algorithms

Finite word length arithmetic roundoff noise in adaptive filter algorithms results in statistical variations in the filter weight vector about the infinite precision arithmetic weight vector. These roundoff errors may be modeled as a statistically non stationary driving noise affecting weight mean and covariance convergence. Mean and covariance expressions and bounds are desired for word lengths in fixed-point arithmetic by making use of multiplication roundoff error models. The adaptive filter algorithms consist of the LMS algorithm, the Widrow-Hoff LMS algorithm, pilot-vector algorithm and clipped vector algorithm. All of these algorithms can be implemented on-line and real-time. However, only the behavior of the LMS algorithm is reported here. The implementation of the adaptive filter algorithms in finite word length arithmetic is most evident in minicomputer, microprocessor, and dedicated digital signal processors for on-line real-time signal identification and parameter estimation in many disciplines. Radar signal processing, adaptive beam forming, acoustic signal identification, communication channel enhancement have a definite need for advanced filtering concepts. Our adaptive algorithms are typically employed in these filter configurations. These filters can also be employed in phase distortion equalizers. A particular advantage of these filters is that they can be trained to equalize a variety of distortions. Should a particular distortion scenario change in time, the filters can be made to easily adapt to the new problem.     

Floorplan design of VLSI circuits

In this paper we present two algorithms for the floorplan design problem. The algorithms are quite similar in spirit. They both use Polish expressions to represent floorplans and employ the search method of simulated annealing. The first algorithm is for the case where all modules are rectangular, and the second one is for the case where the modules are either rectangular or L-shaped. Our algorithms consider simultaneously the interconnection information as well as the area and shape information for the modules. Experimental results indicate that our algorithms perform well for many test problems.     

Random binary search tree with equal elements

We consider random binary search trees when the input consists of a multiset, i.e. a set with multiple occurrences of equal elements, and prove that the randomized insertion and deletion algorithms given by Martínez and Roura (1998) [4] produce random search trees regardless of multiplicities; even if all the elements are equal during the tree updates, a search tree will maintain its randomness. Thus, equal elements do not degenerate a random search tree and they need not to be handled in any special way. We consider also stability of a search tree with respect to its inorder traversal and prove that the algorithms used produce stable trees. This implies an implicit indexing of equal elements giving another proof that multiplicities do not pose problems when maintaining random binary search trees.     

Search heuristics for a parallel machine scheduling problem with ready times and due dates

We consider a problem of scheduling orders on identical parallel machines. An order can be released after a given ready time and must be completed before its due date. An order is split into multiple jobs (batches) and a job is processed on one of the parallel machines. The objective of the scheduling problem is to minimize the holding costs of orders including work-in-process as well as finished job inventories. We suggest two local search heuristics, simulated annealing and taboo search algorithms, for the problem. Performance of the suggested algorithms is tested through computational experiments on randomly generated test problems.     

Occupancy estimation from environmental parameters using wrapper and hybrid feature selection

Occupancy information is essential to facilitate demand-driven operations of air-conditioning and mechanical ventilation (ACMV) systems. Environmental sensors are increasingly being explored as cost effective and non-intrusive means to obtain the occupancy information. This requires the extraction and selection of useful features from the sensor data. In past works, feature selection has generally been implemented using filter-based approaches. In this work, we introduce the use of wrapper and hybrid feature selection for better occupancy estimation. To achieve a fast computation time, we introduce a ranking-based incremental search in our algorithms, which is more efficient than the exhaustive search used in past works. For wrapper feature selection, we propose the WRANK-ELM, which searches an ordered list of features using the extreme learning machine (ELM) classifier. For hybrid feature selection, we propose the RIG-ELM, which is a filter–wrapper hybrid that uses the relative information gain (RIG) criterion for feature ranking and the ELM for the incremental search. We present experimental results in an office space with a multi-sensory network to validate the proposed algorithms.     

Exact and Heuristic Algorithms for Risk‐Aware Stochastic Physical Search

We consider an intelligent agent seeking to obtain an item from one of several physical locations, where the cost to obtain the item at each location is stochastic. We study risk‐aware stochastic physical search (RA‐SPS), where both the cost to travel and the cost to obtain the item are taken from the same budget and where the objective is to maximize the probability of success while minimizing the required budget. This type of problem models many task‐planning scenarios, such as space exploration, shopping, or surveillance. In these types of scenarios, the actual cost of completing an objective at a location may only be revealed when an agent physically arrives at the location, and the agent may need to use a single resource to both search for and acquire the item of interest. We present exact and heuristic algorithms for solving RA‐SPS problems on complete metric graphs. We first formulate the problem as mixed integer linear programming problem. We then develop custom branch and bound algorithms that result in a dramatic reduction in computation time. Using these algorithms, we generate empirical insights into the hardness landscape of the RA‐SPS problem and compare the performance of several heuristics.     

Computational enhancements to fluence map optimization for total marrow irradiation using IMRT

The fluence map optimization (FMO) problem is a core problem in intensity modulated radiation therapy (IMRT) treatment planning. Although it has been studied extensively for site-specific treatment planning, few studies have examined efficient computational methods for solving it for intensity modulated total marrow irradiation (IM-TMI) planning; few studies have also looked at exploiting prior beamlet information to solve the FMO problem in a beam orientation optimization context. In this study, we consider different types of line search strategies and different types of warm-start techniques to improve the speed with which the FMO problem for IM-TMI is solved and the quality of the end solution. We also consider a parallelism-enhanced algorithm to solve the FMO problem for IM-TMI treatment planning with a large number of beams (36 equispaced beams at each of 11 isocenters, for a total of 396 beams). We show that the backtracking line search strategy with step reduction exhibits the best performance and that using either of the two types of warm-start techniques which we consider leads to significant improvements in both solution time and quality. We also provide results for the aforementioned 396-beam plan and show that 30-beam solutions obtained using beam orientation optimization attain a comparable level of quality as this larger solution.     

On the correspondence between AVL trees and brother trees

We discuss the relationships of the classes of height-balanced (search) trees and the classes of brother (search) trees. In particular we characterize each class of height-balanced trees in terms of the class of “corresponding” brother trees and vice versa. Secondly, we show how this characterization leads to the notion of nonstandard updating algorithms. We derive a nonstandard insertion algorithm for height-balanced search trees to illustrate the notion. Finally we consider something of the similarities and differences between the standard and nonstandard insertion algorithms for height-balanced search rees.