Theoretical Algorithm Research Articles

L 2-L ∞ state estimation for continuous stochastic delayed neural networks via memory event-triggering strategy

In this paper, we study the memory event-triggering - state estimation for a type of continuous stochastic neural networks (NNs) subject to time-varying delays. The information of some recent released packets is made use in the proposed triggering conditions to schedule the data propagation, thereby reducing communication frequency and saving energy. By taking into account network-induced complexities (i.e. transmission delays and random disturbances), we first formulate the evolutions of estimation error in an augmented form, and then propose the conditions under with the design goals could be met. By using certain novel Lyapunov–Krasovskii (L–K) functionals in combination with stochastic analysis technique, sufficient conditions have been provided for the existence of desired estimator, guaranteeing both the globally asymptotically mean-square stability and the prescribed - performance simultaneously. Moreover, the estimator gains are obtained by virtue of certain convex optimisation algorithms. Finally, we use an illustrative example to verify the obtained theoretical algorithm.

52 Refinement of a Histologic Algorithm for Burn Depth Categorization Using 1142 Consecutive Burn Wound Biopsies

IntroductionOur group previously reported a theoretical burn biopsy algorithm (BBA-V1) for the categorization of burn wound depth based on histologic analysis, and informed it with the largest series of burn wound biopsies in the literature. That iteration of the BBA resulted in clinical misclassification rates consistent with past literature. Since our last report of that process, we have refined the algorithm with new criteria and a larger repository of burn wound biopsies. Here, we sought to promulgate this newer, simpler version of the BBA (BBA-V2).MethodsThis was an IRB-approved, prospective, multicenter study. Patients with burn wounds assessed by burn experts as requiring excision and autograft underwent 4mm biopsies procured every 25cm2. Serial still photos were obtained at enrollment and at excision intraoperatively.Using H&E with whole slide scanning, a board-certified dermatopathologist assessed each burn biopsy. The criteria used for categorization of burn wound depth in BBA-V1 were: 1) proportion of necrotic adnexal structures, and 2) presence/absence of each of epidermis, papillary dermis, and reticular dermis. The criteria used for BBA-V2 were: 1) magnitude of reticular dermal degeneration, 2) proportion of necrotic adnexal structures, and 3) magnitude of vessel thrombosis.Biopsy pathology results were correlated with still photos by 3 burn experts for consensus of final burn depth diagnosis. Superficial partial thickness (SPT) wounds were considered to be burn wounds likely to have healed without surgery, while deep partial thickness (DPT) and full thickness (FT) were considered unlikely to heal by 21 days.ResultsThe development of BBA V-1 was previously informed by 66 subjects with 117 wounds and 816 biopsies, and resulted in wound categorizations as follows: SPT (20%), DPT (43%), and FT (37%). Therefore, according to BBA-V1, 20% of burn wounds were incorrectly judged as needing excision and grafting by the clinical team. The overall cohort was enlarged to 162 subjects with 294 wounds and 1142 biopsies. The most recent 838 burn wound biopsies were then re-reviewed and re-categorized according to the new BBA-V2 criteria and algorithm. Under BBA-V2, 3% of all burn wound biopsies were categorized as superficial partial thickness, 69% were categorized as deep partial thickness, and 29% were categorized as full thickness.ConclusionsOur study demonstrates that by adding dermal degeneration severity and vessel thrombosis to our previous criterion of adnexal structure necrosis, BBA-V2 had a much higher rate of concordance with visual clinical assessment for burn wounds clinically judged as needing surgical excision. This study serves as the largest analysis of burn biopsies by modern day burn experts.

A large experimentation to analyze the effects of implementation bugs in machine learning algorithms

In the last years, Machine Learning (ML) has become extremely used in software systems: it is applied in many different contexts such as medicine, bioinformatics, finance, automotive, only to mention a few. One of the main drawbacks recognized in the literature is that there are still no consolidated approaches and strategies to ensure the reliability of the code implementing the underlying ML theoretical algorithms. This fact has potentially a strong impact since many critical software systems rely on ML algorithms for implementing intelligent behaviors, and so on (potentially) unreliable code that could cause, in extreme cases, catastrophic errors: e.g., loss of life due to a wrong diagnosis of an ML-based cancer classifier.Our work aims to understand better the impact that implementation bugs have on the results provided by ML algorithms. Such analysis is fundamental to define novel techniques able to detect bugs in ML-based software systems. Thus, we extensively analyzed thousands of bugs on eight ML algorithms (in particular, four classification algorithms and four clustering ones). The bugs were injected by using an automatic Mutation tool able to mimic realistic errors in the algorithms source code.The empirical study shows that a large amount of the injected bugs are silent since they do not influence the results the eight algorithms provide on the 17 datasets employed in our study; in the remaining cases, the bugs emerge as runtime errors, exceptions, or modified accuracy of the predictions. Moreover, we also discovered that about 1% of the injected bugs are extremely dangerous since they drastically affect the quality of the predictions only in rare cases and with specific datasets increasing the possibility of going unnoticed. The fact that a considerable portion of the bugs does not influence the behavior of the algorithms, on the datasets employed in our study, poses a considerable problem: indeed, among them, several other dangerous silent bugs could be present. They could emerge when the implementations of the algorithms are employed on a novel dataset and with different settings. So the problem of the dangerous silent bugs can potentially be more pervasive than shown in our study.

Contrasting accuracies of single and ensemble models for predicting solar and thermal performances of traditional vaulted roofs

Traditional curved-roof forms have significant potentials in mitigating undesirable environmental impacts. Their performance predictions can be grouped into 4 trendlines of varying degrees of sophistication: theoretical abstracts, numerical methods, white-box simulations and black-box machine learning algorithms. Unprecedently, this research investigates the potential contribution of single- and ensemble-models to approximate the average hourly direct normal and diffuse horizontal irradiances (AHIRDirect, AHIRDiffuse) and cooling energy consumption (AHECCooling) of buildings topped with vaulted-roof forms of various configurations in Aswan, Egypt. Solar and energy simulations are first conducted to build essential datasets, which get pre-processed, before developing 8 single-models, representing 4 families of supervised single-algorithms: artificial neural networks, random forests, k-nearest neighbors and support vector regression. Voting ensemble-model is then created by combining the best-performing single-models. Lastly, the accuracies of all models are compared against simulation outputs. The results showed that no single-model could dominantly predict AHIRDirect, AHIRDiffuse and AHECCooling, obtaining tolerable R2 values, ranging from 97.017 to 61.913%, 92.782 to 43.986% and 99.341 to −9.219%, corresponding to RMSE values of 47.321 to 195.208, 17.457 to 53.617 and 0.002 to 0.032, respectively. Alternatively, voting ensemble-model acquired even better R2 values of 93.971, 93.047 and 97.276%, with RMSE values of 69.000, 17.249 and 0.004, respectively.

Computing Essential Sets for Convex and Nonconvex Scenario Problems: Theory and Application

The scenario approach is a general data-driven algorithm to chance-constrained optimization. It seeks the optimal solution that is feasible to a carefully chosen number of scenarios. A crucial step in the scenario approach is to compute the cardinality of essential sets, which is the smallest subset of scenarios that determine the optimal solution. This article addresses the challenge of efficiently identifying essential sets. For convex problems, we demonstrate that the sparsest dual solution of the scenario problem could pinpoint the essential set. For nonconvex problems, we show that two simple algorithms return the essential set when the scenario problem is nondegenerate. Finally, we illustrate the theoretical results and computational algorithms on security-constrained unit commitment (SCUC) in power systems. In particular, case studies of chance-constrained SCUC are performed in the IEEE 118-bus system. Numerical results suggest that the scenario approach could be an attractive solution to practical power system applications.

Generalized Submodular Information Measures: Theoretical Properties, Examples, Optimization Algorithms, and Applications

Generalized Submodular Information Measures: Theoretical Properties, Examples, Optimization Algorithms, and Applications

Towards improving the global water application uniformity of centre pivots through lateral speed adjustment

An initial step is made towards improving the global water application uniformity of centre pivots by adjusting the lateral speed based on diurnal changes of wind drift and evaporation losses (WDEL). A large WDEL dataset was used to construct an artificial neural network (ANN) model capable of predicting the above ground application efficiency (AGAE ≈ 1- WDEL) from input weather data. Given its reasonably good accuracy (RMSE = 1.5%), the ANN model was then applied in heuristic, simple and dynamic algorithms that used forecasted or real-time weather parameters corresponding to the hourly time intervals of the pivot's revolution time. For each timing interval, the final output of the algorithms is a coefficient ( λ ), that when multiplied by the lateral speed, can ensure a more consistent water delivery depth across the field. The proposed algorithms were theoretically tested for extremely hot, dry, humid, and windy days. These efforts indicated that under the climatic conditions of the study, λ could potentially vary between 0.92% and 1.06% if the revolution time is about 24 h. The λ coefficient typically increased from 1:00 AM to 9:00 AM, then decreased until the sunset, between 19:00 to 21:00, and then increased again until midnight. This general trend for λ could however, be disrupted by sudden changes in the wind speed. Future studies should conduct extensive field work to test the efficacy of the proposed algorithms and quantify the improvement of centre pivot global water application uniformity through the lateral speed adjustment. • Centre pivot lateral speed can be adjusted to account for a changing application efficiency. • Two theoretical algorithms were developed to improve the global water application uniformity. • The first algorithm uses forecasted weather data while the second uses real time data. • The lateral speed adjustment factor varied within ±6% during a 24-hr cycle. • Centre pivot lateral speed adjustment is easy to implement.

NucPosDB: a database of nucleosome positioning in vivo and nucleosomics of cell-free DNA

Nucleosome positioning is involved in many gene regulatory processes happening in the cell, and it may change as cells differentiate or respond to the changing microenvironment in a healthy or diseased organism. One important implication of nucleosome positioning in clinical epigenetics is its use in the “nucleosomics” analysis of cell-free DNA (cfDNA) for the purpose of patient diagnostics in liquid biopsies. The rationale for this is that the apoptotic nucleases that digest chromatin of the dying cells mostly cut DNA between nucleosomes. Thus, the short pieces of DNA in body fluids reflect the positions of nucleosomes in the cells of origin. Here, we report a systematic nucleosomics database — NucPosDB — curating published nucleosome positioning datasets in vivo as well as datasets of sequenced cell-free DNA (cfDNA) that reflect nucleosome positioning in situ in the cells of origin. Users can select subsets of the database by a number of criteria and then obtain raw or processed data. NucPosDB also reports the originally determined regions with stable nucleosome occupancy across several individuals with a given condition. An additional section provides a catalogue of computational tools for the analysis of nucleosome positioning or cfDNA experiments and theoretical algorithms for the prediction of nucleosome positioning preferences from DNA sequence. We provide an overview of the field, describe the structure of the database in this context, and demonstrate data variability using examples of different medical conditions. NucPosDB is useful both for the analysis of fundamental gene regulation processes and the training of computational models for patient diagnostics based on cfDNA. The database currently curates ~ 400 publications on nucleosome positioning in cell lines and in situ as well as cfDNA from > 10,000 patients and healthy volunteers. For open-access cfDNA datasets as well as key MNase-seq datasets in human cells, NucPosDB allows downloading processed mapped data in addition to the regions with stable nucleosome occupancy. NucPosDB is available at https://generegulation.org/nucposdb/.

Inverse Mixed Integer Optimization: Polyhedral Insights and Trust Region Methods

Inverse optimization—determining parameters of an optimization problem that render a given solution optimal—has received increasing attention in recent years. Although significant inverse optimization literature exists for convex optimization problems, there have been few advances for discrete problems, despite the ubiquity of applications that fundamentally rely on discrete decision making. In this paper, we present a new set of theoretical insights and algorithms for the general class of inverse mixed integer linear optimization problems. Specifically, a general characterization of optimality conditions is established and leveraged to design new cutting plane solution algorithms. Through an extensive set of computational experiments, we show that our methods provide substantial improvements over existing methods in solving the largest and most difficult instances to date.

On The Neutrosophic Formula of Some Matrix Equations Derived from Data Mining Theory and Control Systems

This paper is dedicated to studying the neutrosophic formula of some famous matrix equations used in theoretical data mining algorithms and control systems by using neutrosophic matrices and refined neutrosophic matrices over neutrosophic real fields. On the other hand, we concentrate on the neutrosophic formula of the Sylvester equation, and Lyapunov equation, where we study their formulas and properties in terms of theorems in the neutrosophic real number field and refined real number field. Also, we illustrate many different examples to clarify the validity of our work.

Simple Paths and Cycles of Directed Graph for Stock Trading Network Based on STP

The stock trading relationship of buyers and sellers in the stock market can be characterized by a directed graph. It is an important way to study stock trading network through simple directed paths and cycles. In the present paper, we establish a seeking model of simple directed paths and cycles, and obtain some necessary and sufficient conditions for simple directed path and cycle in the directed graph. The algorithms of finding simple directed path and cycle of any specified length are given. The main approach we used is the semi‐tensor product of matrices, which can reduce the search space. An illustrative example is given to show that the theoretical results and algorithms are effective.

Theoretical analysis and algorithm design of optimized pilot for downlink channel estimation in massive MIMO systems based on compressed sensing

Aiming at the pilot design problem in channel estimation of large-scale multiple input multiple output (MIMO) systems, an adaptive autocorrelation matrix reduction parameter pilot optimization algorithm based on channel reconstruction error rate minimization is proposed under the framework of compression perception theory. Firstly, the system model and orthogonal matching pursuit (OMP) algorithm are introduced. Secondly, for minimizing the channel reconstruction error rate, the relation between the expected value of the correlation decision in each iteration of the OMP algorithm and the reconstruction error rate is analyzed. For the optimal expected value of the correlation decision, the relation between the channel reconstruction error rate and the correlation of the pilot matrix column under the OMP algorithm is derived, and the two criteria of optimizing the pilot matrix are obtained: the pilot matrix column correlation expectation and the variance minimization. Then the method of optimizing the pilot matrix is studied, and the corresponding adaptive autocorrelation matrix reduction parameter pilot matrix optimization algorithm is proposed. In each iteration, whether the average column correlation degree of the matrix to be optimized is reduced is used as a judgment condition. The autocorrelation matrix reduction parameter value is adjusted to make the parameters close to the theoretical optimization. The simulation results show that the proposed method has a better column correlation property and lower channel reconstruction error rate than the pilot matrix obtained, separately, by Gaussian matrix, Elad method and low power average column correlation method.

Динамическое программирование в задаче маршрутизации: декомпозиционный вариант

The questions of applying the dynamic programming (DP) apparatus to the routing problem with constraints and cost functions with the tasks list dependence are investigated. It is supposed that binary partition of the task set is given; tasks of the first task group must be fulfilled before the fulfillment of the task of the second group begins. In each of the groups, precedence conditions may be present. This setting can be applied in the case of sheet cutting on CNC machines, where two above-mentioned groups form zones planned at the cutting stage. In general case, for the optimal solution construction, the two-stage variant of DP is used. Linking two versions of DP is realized by identification of the criterion terminal component for service problem of the first group with extremum function connected with the second group. The connection of optimal solutions for above-mentioned two problems allows to construct an optimal solution for the initial joint problem. Based on the theoretical constructions algorithm realized on personal computer is constructed; computing experiment is realized.

Data compression algorithms for sensor networks with periodic transmission schemes

The operating state of switch cabinet is significant for the reliability of the whole power system, collecting and monitoring its data through the wireless sensor network is an effective method to avoid accidents. This paper proposes a data compression method based on periodic transmission model under the condition of limited energy consumption and memory space resources in the complex environment of switch cabinet sensor networks. Then, the proposed method is rigorously and intuitively shown by theoretical derivation and algorithm flow chart. Finally, numerical simulations are carried out and compared with the original data. The comparisons of compression ratio and error results indicate that the improved algorithm has a better effect on the periodic sensing data with interference and can make sure the change trend of data by making certain timing sequence.

Inferring network structures via signal Lasso

Inferring the connectivity structure of networked systems from data is an extremely important task in many areas of science. Most real-world networks exhibit sparsely connected topologies, with links between nodes that in some cases may even be associated with a binary state (0 or 1, denoting, respectively, the absence or presence of a connection). Such un-weighted topologies are elusive to classical reconstruction methods such as Lasso or compressed sensing techniques. We here introduce an approach called signal Lasso, in which the estimation of the signal parameter is subjected to 0 or 1 values. The theoretical properties and algorithm of the proposed method are studied in detail. Applications of the method are illustrated for an evolutionary game and synchronization dynamics in several synthetic and empirical networks, for which we show that our strategy is reliable and robust and outperforms the classical approaches in terms of accuracy and mean square errors.

Influence of optical nonlinearity on combining efficiency in ultrashort pulse fiber laser coherent combining system

The influence of optical nonlinearity on combining efficiency in ultrashort pulse fiber laser coherent combining system is investigated theoretically and experimentally. In the theoretical work, a new theoretical algorithm is presented for the coherent combining efficiency, which can be used to quantify the spectral coherence decay induced by optical nonlinearity imbalance between the sub-beams. The spectral information of the sub-beam is obtained by numerically solving the nonlinear Schrödinger equation (NLSE) in this algorithm to ensure an accurate prediction. In the experimental work, the coherent combining of two all-fiber picosecond lasers is achieved, and the influence of imbalanced optical nonlinearity on the combining efficiency is studied, which agrees with the theoretical prediction. This paper reveals the physical mechanism for the influence of optical nonlinearity on the combining efficiency, which is valuable for the coherent combining of ultrashort pulse fiber laser beams.

Visual detection and tracking algorithms for minimally invasive surgical instruments: A comprehensive review of the state-of-the-art

Minimally invasive surgical instrument visual detection and tracking is one of the core algorithms of minimally invasive surgical robots. With the development of machine vision and robotics, related technologies such as virtual reality, three-dimensional reconstruction, path planning, and human–machine collaboration can be applied to surgical operations to assist clinicians or use surgical robots to complete clinical operations. The minimally invasive surgical instrument vision detection and tracking algorithm analyzes the image transmitted by the surgical robot endoscope, extracting the position of the surgical instrument tip in the image, so as to provide the surgical navigation. This technology can greatly improve the accuracy and success rate of surgical operations. The purpose of this paper is to further study the visual detection and tracking technology of minimally invasive surgical instruments, summarize the existing research results, and apply it to the surgical robot project. By reading the literature, the author summarized the theoretical basis and related algorithms of this technology in recent years. Finally, the author compares the accuracy, speed and application scenario of each algorithm, and analyzes the advantages and disadvantages of each algorithm. The papers included in the review were selected through Web of Science, Google Scholar, PubMed and CNKI searches using the keywords: “object detection”, “object tracking”, “surgical tool detection”, “surgical tool tracking”, “surgical instrument detection” and “surgical instrument tracking” limiting results to the year range 1985–2021. Our study shows that this technology will have a great development prospect in the aspects of accuracy and real-time improvement in the future.

Neutronics Calculation Advances at Los Alamos: Manhattan Project to Monte Carlo

The history and advances of neutronics calculations at Los Alamos during the Manhattan Project through the present are reviewed. Substantial improvements to neutron diffusion methods and the invention of both the Monte Carlo neutron transport methods in 1947 and deterministic discrete ordinates Sn in 1953 were all made at Los Alamos just after the Manhattan Project. We briefly summarize early simpler and more approximate neutronics methods and then describe the need to better predict neutronics behavior through consideration of theoretical equations, models and algorithms, experimental measurements, and available computing capabilities and their limitations. This paper briefly covers key advances in deterministic methods during the Manhattan Project. These capabilities, coupled with increasing postwar defense needs and the invention of electronic computing with the Electronic Numeric Integrator and Computer, known as ENIAC, and the Mathematical Analyzer Numerical Integrator and Automatic Computer Model, known as MANIAC, led to the creation of Monte Carlo and deterministic discrete ordinates neutronics transport methods. We note the important role that the scientific comradery between the Los Alamos scientists played in the process. This paper briefly covers the early methods, algorithms, computers, and electronic and women pioneers that enabled Monte Carlo to spread to all areas of science. We focus heavily on these early developments and the subsequent creation of the MCNP® code, advances in its associated nuclear data, and its applications to problems of national defense at Los Alamos.

IBOA-Based Optimization of Cross-Sectional Dimension of Rods for a 3-RRR PPM to Minimize Energy Consumption

For obtaining optimal cross-sectional dimensions of rods for a 3-RRR planar parallel manipulator (PPM) to minimize energy consumption, the inverse dynamics of the manipulator is modeled based on the Newton–Euler method, after which the coefficient matrix of the inverse dynamics equation is decomposed based on matrix theory. Hence, the objective function, that is, the logical relationship between the energy consumption of the manipulator and the cross-sectional dimension of each rod, is established. However, in solving the multidimensional constrained single-object optimization problem, there are difficulties such as the penalty function’s sensitivity to the penalty factors if the problem is transformed into the one of unconstrained multiobjective optimization. Therefore, to properly handle the constraints, an improved butterfly optimization algorithm (IBOA) is presented to ensure that the new iterated point always falls into the feasible region according to the butterfly optimization algorithm (BOA). Finally, the comparisons among the IBOA, particle swarm optimization (PSO), and BOA and further experiments of the physical prototype are implemented to validate the effectiveness of the proposed theoretical model and numerical algorithm. Results indicate that the proposed IBOA is more suitable for solving the constrained single-object optimization problem with better convergence speed and accuracy.

A Double Sky-Hook Algorithm for Improving Road-Holding Property in Semi-Active Suspension Systems for Application to In-Wheel Motor

This paper presents a double sky-hook algorithm for controlling semi-active suspension systems in order to improve road-holding property for application in an in-wheel motor. The main disadvantage of the in-wheel motor is the increase in unsprung masses, which increases after shaking of the wheel, so it has poor road-holding that the conventional theoretical sky-hook algorithm cannot achieve. The double sky-hook algorithm uses a combination of damper coefficients, one from the chassis motion and the other from the wheel motion. Computer simulations using a quarter and full car dynamic models with the road conditions specified by ISO2631 showed the effectiveness of the algorithm. It was observed that the algorithm was the most effective in the vicinity of the wheel hop frequency. This paper also proposed the parameter set of the double sky-hook algorithm to differentiate the driving mode of vehicles under advanced development.