The Least Basic Operations on Heap and Improved Heapsort

The best algorithms of INSERT and DELETE operations on heap is presented,by which HEAPSORT is improved.Inserting on element into and deleting one element from a heap of n elements spend at most [log log n] comparisons and [log n] comparisons and transfers of element in the worst cases respectively.The improved HEAPSORT spends n log n n log log n O(n) comparisons and element transfers (not swap!) in the worst case.It may be the best HEAPSORT algorithm since the lower bound of sorting algorithm [log n!]≈n log n O(n).Especially,in element transfer,this is the best result we known so far.     

An Efficient Adaptive Algorithm for Constructing the Convex Differences Tree of a Simple Polygon

The convex differences tree (CDT) representation of a simple polygon is useful in computer graphics, computer vision, computer aided design and robotics. The root of the tree contains the convex hull of the polygon and there is a child node recursively representing every connectivity component of the set difference between the convex hull and the polygon. We give an O(n log K + K log² n) time algorithm for constructing the CDT, where n is the number of polygon vertices and K is the number of nodes in the CDT. The algorithm is adaptive to a complexity measure defined on its output while still being worst case efficient. For simply shaped polygons, where K is a constant, the algorithm is linear. In the worst case K = O(n) and the complexity is O(n log² n). We also give an O(n log n) algorithm which is an application of the recently introduced compact interval tree data structure.     

改进的二分法查找

当前有很多的查找算法,其中在对有序数列的查找算法中二分法查找（binary search）是最常用的。利用二分法,在含有n个元素的有序数列中查找一个元素的最大比较次数为[logn]＋1。在很多情况中,在查找之前有序数列分布的很多信息为已知,比如说如果知道了有序数列中每相邻两个元素之差的最大值的一个上界,就可以有比二分法更加有效的查找算法。文章给出了一个称之为改进的二分法查找算法。改进的二分法查找性能明显优于二分法查找,受数列分布的影响,其最坏情况下查找一个元素的最大比较次数在1和[logn]＋1之间,明显优于二分查找的[logn]＋1。在实际应用中利用改进的二分法可以极大地提高查找效率。     

一种快速相容三角剖分算法

提出了一种基于凹多边形凸分解的相容三角剖分方法。先将凹边形分解成凸多边形,再对子多边形进行三角剖分,即可实现相容三角剖分。在最坏的情况下添加O（jk)个辅助点,时间复杂度为O(jn+n log n+jk log n)     

超平面覆盖问题的参数化改进算法

超平面覆盖问题是计算几何领域中一类典型的NP难问题,在实际生活中有着广泛的应用.针对NP难问题的难解性,人们提出了一些传统的方法用来求解这些NP难问题.但由于这些方法具有各自的局限性,不能满足实际应用中的各种需求,人们从新的理论角度为固定参数可解的NP难问题设计参数算法.通过深入分析直线覆盖问题(超平面覆盖问题的一个特例)的结构特征,并利用深度有界搜索树的方法,提出了一个时间复杂度为O(k3(0.736k)k+nlogk)的确定性参数算法,极大地改进了当前最好的结果O((k/2.2)2k+nlogk).通过对上述算法在高维空间中的进一步扩展,提出了关于超平面覆盖问题时间复杂度为O(dkd+1(dk)!/((d!)kk!)+nd+1)确定性参数算法,对当前的最好结果O(kd(k+1)+nd+1)有较大改进.     

Halfplanar range search in linear space and O(n0.695) query time

Let S denote a set of n points in the Euclidean plane. A halfplanar range query specifies a halfplane h and requires the determination of the number of points in S which are contained in h. A new data structure is described which stores S in O(n) space and allows us to answer a halfplanar range query in O(n^log₂(1+√5)−1) time in the worst case, thus improving the best result known before. The structure can be built in O(n log n) time.     

An algorithm for merging meaps

Summary We present an algorithm to merge priority queues organized as heaps. The worst case number of comparisons required to merge two heaps of sizes k and n is O(log(n)*log(k)). The algorithm requires O(k) +log(n)*log (k)) data movements if heaps are implemented using arrays and O(log(n)*log(k)) for a pointer-based implementation. Previous algorithms require either O(n+k) data movements and comparisons, or O(k*log(log(n+k))) comparisons and O(k*log(n+k)) data movements. The algorithm presented in this paper improves on the previous algorithms for the case when k>log(n).This work was done while the authors were at McGill University, Montréal, Canada     

A simple Voronoi diagram algorithm for a reconfigurable mesh

In this paper, we introduce a simple and efficient algorithm for computing the Voronoi Diagram for n planar points on a reconfigurable mesh of size O(n)×O(n). The algorithm has a worst case running of O(log n log log n) time. The algorithm exploits the O(1) communication diameter of the reconfigurable mesh model to implement efficient load balancing     

Uniform leader election protocols for radio networks

A radio network is a distributed system with no central arbiter, consisting of n radio transceivers, henceforth referred to as stations. We assume that the stations are identical and cannot be distinguished by serial or manufacturing number. The leader election problem asks to designate one of the stations as leader. In this work, we focus on single-channel, single-hop radio networks. We assume that time is slotted and all transmissions occur at slot boundaries. In each time slot, the stations transmit on the channel with some probability until, eventually, one of the stations is declared leader. A leader election protocol is said to be uniform if, in each time slot, every station transmits with the same probability. In a seminal paper, Willard (1986) presented a uniform leader election protocol for single-channel single-hop radio stations terminating in log log n+o(log log n) expected time slots. It was open for more than 15 years whether Willard's protocol featured the same time performance with "high probability." One of our main contributions is to show that, unfortunately, this is not the case. Specifically, we prove that for every parameter f∈e^O(n), in order to ensure termination with probability exceeding 1-1/f, Willard's protocol must take log log n+Ω(√f) time slots. The highlight of this work is a novel uniform leader election protocol that terminates, with probability exceeding 1-1/f, in log log n+o(log log n)+O(log f) time slots. Finally, we provide simulation results that show that our leader election protocol outperforms Willard's protocol in practice     

Unicast in hypercubes with large number of faulty nodes

Unicast in computer/communication networks is a one-to-one communication between a source node s and a destination node t. We propose three algorithms which find a nonfaulty routing path between s and t for unicast in the hypercube with a large number of faulty nodes. Given the n-dimensional hypercube H_n and a set F of faulty nodes, node uϵ H_n is called k-safe if u has at least k nonfaulty neighbors. The H_n is called k-safe if every node of H_n is k-safe. It has been known that for 0⩽k⩽n/2, a k-safe H_n is connected if |F|⩽2^k(n-k)-1. Our first algorithm finds a nonfaulty path of length at most d(s,t)+4 in O(n) time for unicast between 1-safe s and t in the H_n with |F|⩽2n-3, where d(s,t) is the distance between s and t. The second algorithm finds a nonfaulty path of length at most d(s,t)+6 in O(n) time for unicast in the 2-safe H_n with |F|⩽4n-9. The third algorithm finds a nonfaulty path of length at most d(s,t)+O(k²) in time O(|F|+n) for unicast in the k-safe H_n with |F|⩽2^k(n-k)-1 (0⩽k⩽n/2). The time complexities of the algorithms are optimal. We show that in the worst case, the length of the nonfaulty path between s and t in a k-safe H_n with |F|⩽2^k(n-k)-1 is at least d(s,t)+2(k+1) for 0⩽k⩽n/2. This implies that the path lengths found by the algorithms for unicast in the 1-safe and 2-safe hypercubes are optimal     

A tight lower bound for the worst case of Bottom-Up-Heapsort

Bottom-Up-Heapsort is a variant of Heapsort. Its worst-case complexity for the number of comparisons is known to be bounded from above by 3/2n logn+0(n), wheren is the number of elements to be sorted. There is also an example of a heap which needs 5/4n logn-0(n log logn) comparisons. We show in this paper that the upper bound is asymptotically tight, i.e., we prove for largen the existence of heaps which need at least 3/2n logn–O(n log logn) comparisons. This result also proves the old conjecture that the best case for classical Heapsort needs only asymptotic n logn + O(n log logn) comparisons.This work was supported by the ESPRIT II program of the EC under Contract No. 3075 (project ALCOM).     

确定两个任意简单多边形交、并、差的算法

提出了把多边形的边分为奇偶边的新思想，根据输入多边形A，B之间边的拓扑关系，划分A，B边为内边、外边、重叠边3种，揭示A，B与它们的交、并、差之间边的本质联系，进而描述了确定任意两个简单多边形交、并、差算法．算法的时间复杂度为O((n m k)log(n m k))，其中n，m分别是A，B的顶点数，k是两多边形的交点数．算法建立在数学理论基础之上，很好地处理了布尔运算的奇异情形，比如重叠边，边与边相交于边的顶点等情形．本算法易于编程实现。     

Compact Routing Schemes for Dynamic Ring Networks

We consider the problem of routing in an asynchronous dynamically changing ring of processors using schemes that minimize the storage space for the routing information. In general, applying static techniques to a dynamic network would require significant re-computation. Moreover, the known dynamic techniques applied to the ring lead to inefficient schemes. In this paper we introduce a new technique, Dynamic Interval Routing, and we show tradeoffs between the stretch factor, the adaptation cost, and the size of the update messages used by routing schemes based upon it. We give three algorithms for rings of maximum size N: the first two are deterministic, one with adaptation cost zero but worst case stretch factor \lfloor N/2 \rfloor, the other with worst case adaptation cost O(N) update messages of O(log N) bits and stretch factor 1. The third algorithm is randomized, uses update messages of size O(k log N), has adaptation cost O(k), and expected stretch factor 1+1/k, for any integer k 3. All schemes require O(log N) bits per node for the routing information and all messages headers are of O(log N) bits.     

QuickHeapsort: Modifications and Improved Analysis

QuickHeapsort is a combination of Quicksort and Heapsort. We show that the expected number of comparisons for QuickHeapsort is always better than for Quicksort if a usual median-of-constant strategy is used for choosing pivot elements. In order to obtain the result we present a new analysis for QuickHeapsort splitting it into the analysis of the partition-phases and the analysis of the heap-phases. This enables us to consider samples of non-constant size for the pivot selection and leads to better theoretical bounds for the algorithm. Furthermore, we introduce some modifications of QuickHeapsort. We show that for every input the expected number of comparisons is at most \(n\log _{2}n - 0.03n + o(n)\) for the in-place variant. If we allow n extra bits, then we can lower the bound to \( n\log _{2} n -0.997 n+ o (n)\). Thus, spending n extra bits we can save more that 0.96n comparisons if n is large enough. Both estimates improve the previously known results. Moreover, our non-in-place variant does essentially use the same number of comparisons as index based Heapsort variants and Relaxed-Weak-Heapsort which use \( n\log _{2}n -0.9 n+ o (n)\) comparisons in the worst case. However, index based Heapsort variants and Relaxed-Weak-Heapsort require \({\Theta }(n\log n)\) extra bits whereas we need n bits only. Our theoretical results are upper bounds and valid for every input. Our computer experiments show that the gap between our bounds and the actual values on random inputs is small. Moreover, the computer experiments establish QuickHeapsort as competitive with Quicksort in terms of running time.     

A near optimal isosurface extraction algorithm using the span space

Presents the “Near Optimal IsoSurface Extraction” (NOISE) algorithm for rapidly extracting isosurfaces from structured and unstructured grids. Using the span space, a new representation of the underlying domain, we develop an isosurface extraction algorithm with a worst case complexity of o(√n+k) for the search phase, where n is the size of the data set and k is the number of cells intersected by the isosurface. The memory requirement is kept at O(n) while the preprocessing step is O(n log n). We utilize the span space representation as a tool for comparing isosurface extraction methods on structured and unstructured grids. We also present a fast triangulation scheme for generating and displaying unstructured tetrahedral grids     

Matching Algorithms Are Fast in Sparse Random Graphs

We present an improved average case analysis of the maximum cardinality matching problem. We show that in a bipartite or general random graph on n vertices, with high probability every non-maximum matching has an augmenting path of length O(log n). This implies that augmenting path algorithms like the Hopcroft-Karp algorithm for bipartite graphs and the Micali-Vazirani algorithm for general graphs, which have a worst case running time of O(m√n), run in time O(m log n) with high probability, where m is the number of edges in the graph. Motwani proved these results for random graphs when the average degree is at least ln (n) [Average Case Analysis of Algorithms for Matchings and Related Problems, Journal of the ACM, 41(6):1329-1356, 1994]. Our results hold if only the average degree is a large enough constant. At the same time we simplify the analysis of Motwani.     

Learning efficiency of redundant neural networks in Bayesianestimation

This paper proves that the Bayesian stochastic complexity of a layered neural network is asymptotically smaller than that of a regular statistical model if it contains the true distribution. We consider a case when a three-layer perceptron with M input units, H hidden units and N output units is trained to estimate the true distribution represented by the model with H(0) hidden units and prove that the stochastic complexity is asymptotically smaller than (1/2) {H(0 ) (M+N)+R} log n where n is the number of training samples and R is a function of H-H(0), M, and N that is far smaller than the number of redundant parameters. Since the generalization error of Bayesian estimation is equal to the increase of stochastic complexity, it is smaller than (1/2 n) {H(0) (M+N)+R} if it has an asymptotic expansion. Based on the results, the difference between layered neural networks and regular statistical models is discussed from the statistical point of view.     

Multiway merging in parallel

The problem of merging k (k⩾2) sorted lists is considered. We give an optimal parallel algorithm which takes O((n log k/p)+log n) time using p processors on a parallel random access machine that allows concurrent reads and exclusive writes, where n is the total size of the input lists. This algorithm achieves O(log n) time using p=n log k/log n processors. Most of the previous log n research for this problem has been focused on the case when k=2. Very recently, parallel solutions for the case when k=2 have been reported. Our solution is the first logarithmic time optimal parallel algorithm for the problem when k⩾2. It can also be seen as a unified optimal parallel algorithm for sorting and merging. In order to support the algorithm, a new processor assignment strategy is also presented     

A simplified complexity analysis of mcdiarmid and reed's variant of bottom-up-heapsort

McDiarmid and Reed (1989) presented a variant of BOTTOM-UP-HEAPSORT which requires nlog₂ n+n element comparisons (for n= 2 ^h+1-1) in the worst case, but requires an extra storage of n bits. Ingo Wegener (1992) has analyzed the average and worst case complexity of the algorithm which is very complex and long. In this paper we present a simplified complexity analysis of the same algorithm from a different viewpoint. For n= 2 ^h+1-1, we show that it requires nlog₂ n+n element comparisons in the worst case and nlog₂ n+0.42n comparisons on the average     

On the Complexity of Computing the Hypervolume Indicator

The goal of multiobjective optimization is to find a set of best compromise solutions for typically conflicting objectives. Due to the complex nature of most real-life problems, only an approximation to such an optimal set can be obtained within reasonable (computing) time. To compare such approximations, and thereby the performance of multiobjective optimizers providing them, unary quality measures are usually applied. Among these, the hypervolume indicator (or S-metric) is of particular relevance due to its favorable properties. Moreover, this indicator has been successfully integrated into stochastic optimizers, such as evolutionary algorithms, where it serves as a guidance criterion for finding good approximations to the Pareto front. Recent results show that computing the hypervolume indicator can be seen as solving a specialized version of Klee's measure problem. In general, Klee's measure problem can be solved with O(n logn + nd/2logn) comparisons for an input instance of size n in d dimensions; as of this writing, it is unknown whether a lower bound higher than Omega(n log n) can be proven. In this paper, we derive a lower bound of Omega(n log n) for the complexity of computing the hypervolume indicator in any number of dimensions d > 1 by reducing the so-called uniformgap problem to it. For the 3-D case, we also present a matching upper bound of O(n log n) comparisons that is obtained by extending an algorithm for finding the maxima of a point set.