Random Search

The number of descendants of a node in a binary search tree (BST) is the size of the subtree having this node as a root; the number of ascendants is the number of nodes on the path connecting this node with the root. Using a purely combinatorial approach (generating functions and differential equations) we are able to extend previous results. For the number of descendants we get explicit formulaae for all moments; for the number of ascendants, which is harder, we get the variance. A natural extension of binary search trees occurs when performing local reorganisations. Poblete and Munro have already analyzed some aspects of these locally balanced binary search trees (LBSTs). Here, we relate these structures with the performance of median-of-three Quicksort. We get as new results the variances for ascendants and descendants in this setting. If the rank of the node itself is picked at random ("grand averages"), the corresponding parameters only depend on the size n. In this instance, we get all the moments for the descendants (BST and LBST), as well as the probabilities. For ascendants (LBST), we get the variance and (in principle) the higher moments, as well as the (normal) limiting distribution. The emphasis is on explicit formulaae, and these are sometimes quite involved. Thus, in some instances, we have decided to state abridged versions in the paper and collect the long forms into an appendix that can be downloaded from the URLs

The number of descendants of a node in a binary search tree (BST) is the size of the subtree having this node as a root; the number of ascendants is the number of nodes on the path connecting this node with the root. Using a purely combinatorial approach (generating functions and differential equations) we are able to extend previous results. For the number of descendants we get explicit formulæ for all moments; for the number of ascendants, which is harder, we get the variance. A natural extension of binary search trees occurs when performing local reorganisations. Poblete and Munro have already analyzed some aspects of these locally balanced binary search trees (LBSTs). Here, we relate these structures with the performance of median–of–three Quicksort. We get as new results the variances for ascendants and descendants in this setting. If the rank of the node itself is picked at random ("grand averages"), the corresponding parameters only depend on the size $n$. In this i...

A randomized algorithm with stream splitting for design of heat exchanger networks is presented in this work. The algorithm has provisions for splitting any one of the process streams. We have studied three benchmark problems taken from literature. The results obtained from these studies clearly indicate the strength of the randomization method in finding a cost-effective network. This random search method can find better networks, which are sometimes unnoticed by other optimization techniques. The simplicity of this method as well as the networks obtained is an additional attractive feature, which should encourage the designers to use it. From the results of this study, randomization is recommended as a reliable check for designing heat exchanger networks.

In this paper we present a new optimization algorithm based on a model of the foraging behavior of a population of primitive ants (Pachycondyla apicalis). These ants are characterized by a relatively simple but efficient strategy for prey search in which individuals hunt alone and try to cover a given area around their nest. The ant colony search behavior consists of a set of parallel local searches on hunting sites with a sensitivity to successful sites. Also, their nest is periodically moved. Accordingly, the proposed algorithm performs parallel random searches in the neighborhood of points called hunting sites. Hunting sites are created in the neighborhood of a point called nest. At constant intervals of time the nest is moved, which corresponds to a restart operator which re-initializes the parallel searches. We have applied this algorithm, called API, to numerical optimization problems with encouraging results.

As desktop computer computational power continues to increases dramatically, it is becoming commonplace to run a combination of deadline-sensitive applications. Despite the proliferation of computational power, the detailed nature of these applications causes new problems for the system resource allocation mechanisms. First, these applications are designed to meet their deadlines as long as nearly all the system's resources are available to them; once the system approaches saturation, the collective applications will fail to meet their deadlines. To aggravate the situation, conventional best effort managers will allocate resources to the competing applications based on a static form of equitability rather than addressing the dynamic relative benefit provided by each application. Second, the applications differ from conventional real-time applications: though members of this new class of desktop applications are sensitive to deadlines, their constraints are non-critical. They are also typically designed to use the full spectrum of functions provided by a general purpose system call interface rather than the more limited interface of a real-time OS. This paper describes a flexible soft real-time system design that addresses these two problems. The CPU scheduling strategy accommodates the community of applications by taking application benefit into account during times of heavy load. Applications are written to use the full interface of a general purpose system call interface, yet the system is able to schedule them according to their deadlines and resource reservation and availability. This paper describes the theoretical foundation of the approach, additional application responsibilities, the design of a middleware system to implement the approach, and then it presents an extensive set of experimental studies that demonstrate the behavior of the approach. The Dynamic QoS Manager middleware (implemented on top of two different versions of Unix), is shown to be highly effective system for supporting contemporary soft real-time desktop applications.

We study the role of dynamical constraints in the general problem of finding the best statistical strategy for random searching when the targets can be detected only in the limited vicinity of the searcher. We find that the optimal search strategy depends strongly on the delay time during which a previously visited site becomes unavailable. We also find that the optimal search strategy is always described for large step lengths ' by a power-law distribution P' ' ÿ , with 1 < 2. Our findings appear to remain valid even if arbitrary energy costs of locomotion are considered.

Recent developments in Neural Architecture Search (NAS) resort to training the supernet of a predefined search space with weight sharing to speed up architecture evaluation. These include random search schemes, as well as various schemes based on optimization or reinforcement learning, in particular policy gradient, that aim to optimize a parametric architecture distribution and the shared model weights simultaneously. In this paper, we focus on efficiently exploring the important region of a neural architecture search space with reinforcement learning. We propose Deep Deterministic Architecture Sampling (DDAS) based on deep deterministic policy gradient and the actor-critic framework, to selectively sample important architectures in the supernet for training. Through balancing exploitation and exploration, DDAS is designed to combat the disadvantages of prior random supernet warm-up schemes and optimization schemes. Gradient-based NAS approaches require the execution of multiple sh...

An important problem addressed using cDNA microarray data is the detection of genes differentially expressed in two tissues of interest. Currently used approaches ignore the multidimensional structure of the data. However it is well known that correlation among covariates can enhance the ability to detect less pronounced differences. We use the Mahalanobis distance between vectors of gene expressions as a criterion for simultaneously comparing a set of genes and develop an algorithm for maximizing it. To overcome the problem of instability of covariance matrices we propose a new method of combining data from small-scale random search experiments. We show that by utilizing the correlation structure the multivariate method, in addition to the genes found by the onedimensional criteria, finds genes whose differential expression is not detectable marginally.

We consider a supply chain, which consists of N stocking locations and one supplier. The locations may be coordinated through replenishment strategies and lateral transshipments, i.e., transfer of a product among locations at the same echelon level. The supplier has limited production capacity. Therefore, the total amount of product supplied to the N locations is limited in each time period. When total replenishment orders exceed total supply, not all locations will be able to attain their base stock values. Therefore, different allocation rules are considered to specify how the supplier rations its limited capacity among the locations. We team up the modeling flexibility of simulation with sample path optimization to address the multilocation transshipment problem. We solve the sample average approximation problem by random search and by gradient search. With this numerical approach, we can study problems with non-identical costs and correlated demand structures.

In different agro-climates the greenhouse horticulture are making use of various types of plastic cover where its influence on production in relation to crop, indoor climate and outside climate not considered in selection. In this study, major spectral properties of mainly used greenhouse cover for Ethiopian highland and their effect on yield were investigated. Cover spectral properties such as cover near-infrared (NIR) transmission, photosynthetic active radiation (PAR) transmission and cover longwave emission coefficient were major parameters that are considered in this study to determine their effect on dry matter harvest. To examine the influence of the cover parameters on biomass production, a generic tool or a model constructed in Matlab software was used. The model comprises energy balance equations, mass balance equations and growth model. The measured greenhouse indoor temperature data was considered for validation of the model. Simulations of indoor temperature, canopy temperature and vapor concentration were conducted for the existing greenhouse configuration. The result of the simulation showed that in Ethiopian greenhouse for highland, high and low temperature, night time high humidity and low CO 2 during day time are major problems which inhibit the growth and on the other hand create favorable condition for well know disease in the region that is botrytis. According to sensitivity analysis output of the model, 1% increase in PAR transmission resulted 0.009% increase dry matter harvest, whereas a percent increase in cover NIR transmission has caused 0.259% reduction in yield. Moreover, increment of cover long wave emission coefficient by 1 percent positively influences (0.044%) the rose production. The resulted depicted that as compared to low temperature problem during rainy season (2-3 months), high temperature problem during summer season (9-10 months) would become a foremost problem in Ethiopian highland greenhouse horticulture. Thus, plastic cover with lower NIR transmission and higher PAR transmission and long-wave emission coefficient is desirable for Ethiopian highland greenhouses. However, the study revealed that night time low temperature particularly during rainy season should be studied to incorporate in respective growth model.

In Ethiopia almost most greenhouses are equipped with fixed ventilation owing to the fact that its influence on CO2 concentration, indoor temperature and RH% which affects plant growth is not clearly understood or due to lack of capacity. Particularly, relative humidity in the greenhouse not only affects plant growth but also major factor for pest-disease which should be maintained at the required range. In this study, the influences of fixed ventilation opening on indoor CO2 concentration, temperature and relative humidity which consequently affect yield were investigated. To execute simulation and examine the influence of ventilation opening on indoor climate and yield a generic tool or a model set up in Matlab software was employed. For the existing ventilation configuration, simulation of indoor climate was conducted and the result illustrated that low CO2 particularly during day time and high and low temperatures are major problems in the greenhouse. The observed indoor climate...

The paper gives an analysis of some optimization algorithms in computer sciences and their implementation in solving the problem of binary character recognition. The performance of these algorithms is analyzed using the Optimization Algorithm Toolkit, with emphasis on determining the impact of parameter values. It is shown that these types of algorithms have shown a significant sensitivity as far as the choosing of the parameters was concerned which would have an effect on how the algorithm functions. In this sense, the existence of optimal value was found, as well as extremely unacceptable values.

The problem of choosing a good parameter setting for a better generalization performance in a learning task is the so-called model selection. A nested uniform design (UD) methodology is proposed for efficient, robust and automatic model selection for support vector machines (SVMs). The proposed method is applied to select the candidate set of parameter combinations and carry out a k-fold cross-validation to evaluate the generalization performance of each parameter combination. In contrast to conventional exhaustive grid search, this method can be treated as a deterministic analogue of random search. It can dramatically cut down the number of parameter trials and also provide the flexibility to adjust the candidate set size under computational time constraint. The key theoretic advantage of the UD model selection over the grid search is that the UD points are "far more uniform" and "far more space filling" than lattice grid points. The better uniformity and space-filling phenomena make the UD selection scheme more efficient by avoiding wasteful function evaluations of close-by patterns. The proposed method is evaluated on different learning tasks, different datasets as well as different SVM algorithms.

This paper presents a technically advanced, interactive AI-driven interview preparation system designed to simulate realistic interview dialogues. The proposed system dynamically engages candidates by asking follow-up questions, interpreting their responses, and providing personalized feedback. Unlike conventional chatbots with fixed question flows, this model enables bi-directional communication-allowing candidates to both answer and ask questions-thereby closely mirroring genuine interview environments. Leveraging modern NLP technologies, the system generates context-aware interview questions based on industry standards and user preferences.

In this emerging, difficult field, obtaining the best possible patient outcomes depends on the skillful interpretation and punctuation of audiological reports. The Audiogram Digitization Tool (ADT) is a novel method to improve the accessibility and ubiquitous hearing reports for reporting audiograms. For use in electronic health record systems, the ADT transforms conventional audiograms into a standard digital format. Using cutting-edge image processing techniques, it delivers the important hearing threshold data that are useful for data interpretation and correlation with clinical consequences. According to preliminary testing, the ADT can expedite diagnosis by preventing human mistake and reducing data entry time. Additionally, doctors can view data in photos and be informed of any changes in a patient's state, making it easier for experts to collaborate.

These days, a lot of applications, such shortlisting candidates for recruiting processes, depend on information retrieval technologies. This research study presents the use of the robust ranking algorithm Best Match 25 (BM25) for information retrieval in the context of applicant shortlisting. The approach aims to improve the accuracy and efficacy of candidate shortlisting in comparison to earlier methods. This study proposes a method for integrating BM25 into the hiring process to facilitate the selection of qualified candidates from a corpus of resumes or candidate CVs. This study offers a candidate shortlisting system that uses code-driven information retrieval techniques to assist recruiters and HR professionals in locating people with specific skill sets. This code-driven system integrates custom parsers and query processors to efficiently match job requirements with job descriptions. By combining indexing, ranking algorithms, and relevance scoring, the system provides a personalized selection of top candidate matches, streamlining the shortlisting process. This approach is also adaptable and scalable to handle big datasets, which makes it appropriate for a range of hiring processes. This study emphasizes the special role of code-driven solutions in modern candidate shortlisting, stressing the need for efficient information retrieval techniques to satisfy talent acquisition requirements.

The COVID-19 pandemic highlighted the need for efficient and scalable respiratory support systems in emergency and critical care settings. Despite being widely used, manual ventilation bags provide challenges in ensuring consistent airflow, tidal volume, and real-time patient requirements adaption, especially when medical personnel utilize them for prolonged periods of time. The inability of manual Ambu bags to maintain a constant oxygen supply and pressure A new concept for a "Smart Ambu Bag" that combines an automated pressure regulating system with an adjustable oxygen concentration range is presented in this research. By tackling the challenges of manual breathing in emergency scenarios, the method aims to enhance accuracy and patient outcomes. The main objective is to develop a Smart Ventilation Bag with automated pressure adjustments and real-time oxygen monitoring. The Intelligent Ventilation Bag To measure the oxygen concentration in accordance with the patient's requirements, an oxygen sensor is incorporated into the air mix. When compared to existing manual and automated ventilators during COVID-19 situations, the Smart Ventilation Bag has the ability to bridge the gap between accessibility and performance, particularly in locations with limited resources. Better patient outcomes, less caregiver fatigue, and more reliable ventilation supply are demonstrated by simulation and experiment results. The study highlights how the Smart Ventilation Bag can assist expand the capacity for emergency respiratory care, paving the way for more scalable and reasonably priced respiratory support alternatives.

A theoretical and applied literature has suggested that foragers search using Lévy flights, since Lévy flights can maximize the efficiency of search in the absence of information on the location of randomly distributed prey. Foragers, however, often have available to them at least some information about the distribution of prey, gained either through evolved mechanisms, experience and memory, or social transmission of information. As such, we might expect selection for heuristics that make use of such information to further improve the efficiency of random search. Here we present a general model of random search behavior that includes as special cases: area-restricted search, correlated random walks, Brownian search, and Lévy flights. This generative model allows foragers to adjust search parameters based on encounter-conditional and other heuristics. Using a simulation model, we demonstrate the efficiency gains of these search heuristics, and illustrate the resulting differences in...

Slicing has been widely applied in many fields of software engineering, such as debugging, testing, software maintenance and restructuring. Several meta-heuristic search algorithms have been discussed and applied in many areas and been proved to work effectively. Hence, the concept of formulating slicing to search-based applications under the framework of SBSE is arose. This report introduces an approach to locate dependent structures in a program by exploring all the combinations of possible program slices. The paper formulates this problem as a search based software engineering problem. To evaluate the approach, the paper introduces an instance of a search based slicing problem concerned with locating sets of slices that decompose a program into a set of covering slices that minimize inter-slice overlap. The paper reports on algorithm performance and efficiency for implementations of genetic algorithms, random search, hill climbing, simulated annealing and greedy algorithm applied to a set of six C programs. Afterward, a multi-objective genetic algorithm, Niched Pareto Genetic Algorithm, is also introduced into this application. Results from this application show the distribution of the individuals and the tradeoff Pareto front are very positive.

Determinantal point processes (DPPs) are distributions over sets of items that model diversity using kernels. Their applications in machine learning include summary extraction and recommendation systems. Yet, the cost of sampling from a DPP is prohibitive in large-scale applications, which has triggered an effort towards efficient approximate samplers. We build a novel MCMC sampler that combines ideas from combinatorial geometry, linear programming, and Monte Carlo methods to sample from DPPs with a fixed sample cardinality, also called projection DPPs. Our sampler leverages the ability of the hit-and-run MCMC kernel to efficiently move across convex bodies. Previous theoretical results yield a fast mixing time of our chain when targeting a distribution that is close to a projection DPP, but not a DPP in general. Our empirical results demonstrate that this extends to sampling projection DPPs, i.e., our sampler is more sample-efficient than previous approaches which in turn translates to faster convergence when dealing with costly-to-evaluate functions, such as summary extraction in our experiments.

Determinantal point processes (DPPs) are repulsive point processes that have been used for Monte Carlo integration. In this paper, we compare two ways of using DPPs for numerical integration: using the DPP sample directly as quadrature nodes, or using the DPP to construct control variates. We provide theoretical analysis and experimental comparisons of both approaches.

The rapid advancement of Artificial Intelligence (AI) technologies is reshaping the global job market, influencing employment patterns, skill requirements, and workforce dynamics across industries. AI systems ranging from machine learning algorithms to intelligent automation are streamlining operations, augmenting decision-making, and replacing repetitive tasks. While these innovations enhance productivity and economic efficiency, they also raise concerns about job displacement, reskilling needs, and equitable access to emerging opportunities. This review paper examines recent developments in AIdriven labor transformation, evaluates sector-specific impacts, highlights implementation challenges, and explores future directions for inclusive and adaptive workforce strategies.

A system oriented to develop an image processing system, which takes advantage of radiation flashes from a plasma focus, by optimizing the emission-detection-reconstruction procedure, is presented. A computer technique for 3D reconstructions was combined with ...

Abstract—With the growth of the Internet, Internet Service Providers (ISPs) try to meet the increasing traffic demand with new technology and improved utilization of existing resources. Routing of data packets can affect network utilization. Packets are sent along network paths from source to destination following a protocol. Open Shortest Path First (OSPF) is the most commonly used intra-domain Internet routing protocol (IRP). Traffic flow is routed along shortest paths, splitting flow at nodes with several outgoing links on a ...

India's economy is heavily reliant on agriculture, ranking among the top crop producers globally. With farm production contributing approximately 18.1 percent to the nation's GDP, it holds a pivotal role in the agribusiness sector. Challenges such as unproductive farmland and crop diseases hinder annual food production rates. Accurately detecting crop infections is challenging, as inaccuracies could lead to decreased yields and diminished market value. Innovative approaches are thus crucial to facilitate early detection and the development of resilient crops. This study aims to integrate agricultural practices with artificial intelligence to mitigate losses from plant leaf infections. Crops like apples, potatoes, cherries, and grapes exhibit distinct disease symptoms, enabling prompt identification and preventive actions. Employing four models utilizing various CNN architectures-VGG16, VGG19, InceptionV3, and ResNet50-we aimed to identify and predict plant diseases. Through trials on the standard PlantVillage dataset, VGG19 demonstrated superior accuracy compared to other models. We categorized the data into 38 classes for prediction purposes. This proposed model targets agriculture and agro-based businesses, empowering farmers with early intervention strategies to enhance crop productivity.

Recent developments in the field of deep learning models for agricultural pest classification, particularly in relation to oil palm cultivation, highlight the possibility of accurate and effective pest identification. The goal of this study is to use deep learning computation to classify pests in oil palm. The main goal is to evaluate this method's effectiveness in identifying pests. The GoogleNet architecture, GoogleNet with fine-tuned grid search, and GoogleNet with fine-tuned random search are all used in the study. A thorough examination of the performance of the three models is carried out using a variety of assessment metrics, including accuracy, precision, recall, and F1-Score. Photographs of pests on oil palm plants are included in the dataset. While models improved by grid search and random search show significant performance improvement, approaching nearly perfect evaluation metrics, the default GoogleNet model exhibits high accuracy. These results imply that customization improves the model's precision and effectiveness. The study's findings highlight the efficiency of GoogleNetbased models in oil palm farms for classifying pests, with fine-tuning considerably improving their output. In order to advance pest monitoring and management in oil palm cultivation, future research avenues should prioritize dataset expansion, additional model optimization, and the integration of drone-based automatic control and Internet of Things (IoT) technologies.

Two high-pressure insulating phases of lithium were predicted using random search and evolutionary algorithm methods with first-principles electronic structure calculations. It is shown that lithium will transform from the metallic cubic cI16 phase to an insulating monoclinic C2 structure at 74 GPa. The C2 structure is the most stable phase up to 91 GPa, where it transforms to a second insulating orthorhombic Aba2 structure. The C2 and Aba2 structures are the first theoretical models explicitly showing the band gap opening in compressed lithium. The theoretical findings are supported by recent experimental evidence from electrical resistance and x-ray diffraction measurements.

This paper describes the procedure of optimizing the sector geometry of the magnet to obtain the desired isochronous field. The hill shape of the magnet is described in terms of a small number of parameters which are iteratively determined by random search technique geared to minimize the frequency error. 3D magnetic field data and results of equilibrium orbit code are used as input for the iterative optimization process.

In this paper we consider the two-dimensional (2D) rectangular packing problem, where a ®xed set of items have to be allocated on a single object. Two heuristics, which belong to the class of packing procedures that preserve bottomleft (BL) stability, are hybridised with three meta-heuristic algorithms (genetic algorithms (GA), simulated annealing (SA), na õve evolution (NE)) and local search heuristic (hill-climbing). This study compares the hybrid algorithms in terms of solution quality and computation time on a number of packing problems of dierent size. In order to show the eectiveness of the design of the dierent algorithms, their performance is compared to random search (RS) and heuristic packing routines.