Probabilistic Optimization Algorithm Research Articles

MT-RSL: A multitasking-oriented robot skill learning framework based on continuous dynamic movement primitives for improving efficiency and quality in robot-based intelligent operation

Robot skill learning is one of the international advanced directions in the field of robot-based intelligent manufacturing, which makes it possible for robots to learn and operate autonomously in complex real-world environments. In this paper, we propose a multitasking-oriented robot skill learning framework named MT-RSL to improve the efficiency and robustness of multi-task robot skill learning in complex real-world environments, and present the detailed design steps of three key sub-modules included in MT-RSL, namely, sub-task segmentation module, robot skill learning module, and robot configuration optimization module. Firstly, we design a novel sub-task segmentation module based on a coarse-to-fine sub-task segmentation (CF-STS) strategy, in which the Finite State Machine (FSM) is used to analyze complex robot behaviors to obtain a coarse robot sub-task sequence, and the Beta Process Autoregressive Hidden Markov Model (BP-AR-HMM) is used to establish the connection and dependence between multiple demonstration trajectories and encode these trajectories, so as to obtain a finer robot action sequence. Secondly, we extend the basic DMPs system to a continuous dynamic movement primitives (CDMPs) system to construct a novel robot skill learning module, which improves the efficiency of the robot to learn skills and perform actions by orderly coordinating sub-parts such as the activation signals, motion actuator, DMPs-based learning module, and robot configuration optimization module. Then, we design a novel robot configuration optimization module, which introduces the velocity directional manipulability measure (VDM) as the evaluation index of robot kinematic performance to establish the robot configuration optimization model, and proposes an improved probabilistic adaptive particle swarm optimization (Pro-APSO) algorithm to solve this optimization model, so as to obtain the optimal robot configuration. Finally, we develop an experimental testing platform based on the Robot Operating System (ROS) and conduct a series of prototype experiments in complex real-world scenarios. The experimental results demonstrate that our proposed MT-RSL can significantly improve the effectiveness and robustness of multi-task robot skill learning, and can outperform existing robot skill learning methods in terms of both learning efficiency, VDM, and success rate.

Optimizing Enhanced Extended Topological Active Nets Model Using Parallel Processing

In numerous clinical applications that support the diagnosis and treatment planning of a broad variety of disorders, medical image segmentation is essential. Medical picture segmentation using the Enhanced Extended Topological Active Net (EETAN) model has proven to be successful in correctly identifying structures. This study suggests a novel way to combine the best clustering techniques and parallel processing approaches to maximize the segmentation performance of the EETAN model. The Probabilistic Depth Search Optimization (PDSO) Algorithm, which makes the parallel searching technique to find the ideal contour set, is responsible for this. This work implements parallel processing and ideal clustering to improve the EETAN model's performance in medical image segmentation. Performance metrics like accuracy, precision, recall, dice similarity, and computational time are used for a comparison study. The results demonstrate the notable enhancements attained by employing parallel processing and effective clustering.

An unconstrained relaxation digital‐to‐analog converter using optimal bit sequence

AbstractBased on the relaxation digital‐to‐analog converter (ReDAC) architecture and the capacitor charge/discharge function under discrete time conditions, this paper proposes a new mathematical model for the output voltage of the ReDAC to implement the unconstrained relaxation DAC (Uc‐ReDAC). The system clock calibration is not required. An adaptive probabilistic binary particle swarm optimization (AP‐BPSO) algorithm is proposed for solving the binary coding configuration problem under the Uc‐ReDAC model. Simulation results show that the 8‐bit (10‐bit) Uc‐ReDAC has an average sampling rate of 7.9 MS/s (6.76 MS/s).

Improved Bayesian Optimization Framework for Inverse Thermal Conductivity Based on Transient Plane Source Method

In order to reduce the errors caused by the idealization of the conventional analytical model in the transient planar source (TPS) method, a finite element model that more closely represents the actual heat transfer process was constructed. The average error of the established model was controlled at below 1%, which was a significantly better result than for the analytical model, which had an average error of about 5%. Based on probabilistic optimization and heuristic optimization algorithms, an optimization model of the inverse heat transfer problem with partial thermal conductivity differential equation constraints was constructed. A Bayesian optimization algorithm with an adaptive initial population (BOAAIP) was proposed by analyzing the influencing factors of the Bayesian optimization algorithm upon inversion. The improved Bayesian optimization algorithm is not affected by the range and individuals of the initial population, and thus has better adaptability and stability. To further verify its superiority, the Bayesian optimization algorithm was compared with the genetic algorithm. The results show that the inversion accuracy of the two algorithms is around 3% when the thermal conductivity of the material is below 100 Wm-1K-1, and the calculation speed of the improved Bayesian optimization algorithm is three to four times faster than that of the genetic algorithm.

Probabilistic warm solutions-based multi-objective optimization algorithm, application in optimal design of water distribution networks

Single objective optimization of water distribution networks (WDNs) based on construction costs reduces system resiliency. Thus, researchers have developed various optimization frameworks to maximize system resilience while minimizing construction costs during the last decade. They generally utilized population-based meta-heuristic algorithms. These algorithms start with random populations (initial cold solutions), so they need high computational time to converge to the optimal solutions. This paper aims to reduce search space by introducing prebaked initial or warm solutions based on the hydraulic properties of WDNs which are discharge variance and dissipated power function. After generating initial warm solutions, several probabilistic models are used to bring them closer to the global optimal Pareto front and increase their diversity. The proposed approach is used to optimize the design of two well-known benchmarks, Alprovits and Hanoi, and a real-world large-scale WDN. Results show that using warm solutions as the initial population reduces time, the number of iterations, and fitness evaluations to reach a certain accuracy level, more than two times on average for the benchmarks. Also, results show that the proposed method reduces the average construction costs by 45% compared to the traditional optimization method in the same execution time for real-world WDN.

A joint hydrogen and syngas chemical kinetic model optimized by particle swarm optimization

In this work, we propose a novel data-driven framework for detailed kinetic mechanisms optimization applying the heuristic algorithm, namely canonic Particle Swarm Optimization (PSO). The PSO is more effective and robust in coping with uncertainties and incomplete information than deterministic and probabilistic optimization algorithms and is more suitable for machine learning applications. In the proposed framework, to avoid trapping in a local minimum, 1000 local optimums have been obtained and statistically handled to select the final feasible model parameter set with reduced uncertainty intervals and parameter correlations. The developed framework was successfully used for the optimization of the joint H2 and syngas oxidation chemical kinetic model. The data set collected for the model optimization includes 41 reactions and 16 species, and 3000 experimental data targets supplied with uncertainty boundaries measured in shock tubes, jet stirred reactors, plug flow reactors, and premixed laminar flames under wide ranges of temperature, pressure, equivalence ratio, and H2/CO ratios. The initially estimated uncertainties of the reaction rate constants for 15 key reactions were significantly constrained. The reaction rate constants for the H2 oxidation sub-model were re-optimized and their uncertainties were further reduced.

Enhancing static voltage stability of a power system in the presence of large-scale PV plants using a battery energy storage control scheme by the probabilistic technique

This paper aims to investigate the impact of large-scale photovoltaic plant (LS-PVP) controllers on the voltage stability (VS) of a power system. Besides, a new control strategy is proposed for the large-scale battery energy storage system (LS-BESS) to enhance a power system integrated with LS-PVPs, especially VS. To this end, a probabilistic technique, based on the stochastic nature of LS-PVPs and loads, is employed to accurately evaluate the impact of LS-PVPs on the system. Next, a new coupling between the optimization algorithm and the probabilistic technique is proposed to determine the optimal number, location, and capacity of the LS-BESSs by the probabilistic particle swarm optimization (PPSO) algorithm. The objective function is then used to simultaneously maximize the total VS and minimize the total power losses, line loading, as well as investment costs of LS-BESSs. Furthermore, technical performance indices (TPIs) are used for a more comprehensive investigation to compare the proposed solution to the synchronous generators. Finally, several disturbances occurring in the Nordic system are used to validate the effectiveness of the proposed approach, using the Monte Carlo simulation (MCS) method. According to the results, when a fault occurs, a system integrated with LS-PVPs in the stressed area does not operate better than synchronous generators. Therefore, an advanced control method is required to reduce undesirable effects of LS-PVPs. Using the proposed solution, voltage instability was prevented, LS-BESS investment costs were minimized, and negative effects of LS-PVPs on the system were reduced. Besides, costs required for expanding the transmission network as well as communication between controllers and reactive power compensation devices were reduced.

An Intelligent Probabilistic Whale Optimization Algorithm (i-WOA) for Clustering in Vehicular Ad Hoc Networks

An Intelligent Probabilistic Whale Optimization Algorithm (i-WOA) for Clustering in Vehicular Ad Hoc Networks

An uncertainty-aware dynamic shape optimization framework: Gravity dam design

Uncertainties such as material randomness, manufacturing anomalies, and external loading play an important role in the design of engineering structures. Therefore, reliability-based design optimization (RBDO) is frequently used as a tool to guarantee economic aspects without compromising safety. In this paper, an uncertainty-aware framework is proposed for seismic optimization of structures. This algorithm is founded on the stochastic dynamic stressor, and therefore, reduces the bias due to the aleatory nature of ground motion. Safety constraints are evaluated in quantiles instead of the exact solution of reliability analysis. Kriging approach approximates the actual model, while a global search algorithm solves the RBDO problem. The proposed algorithm is used for the optimal shape design of gravity dams with a series of local and global time-variant/invariant performance indices. Two sets of deterministic and probabilistic shape optimization algorithms were compared to demonstrate the impact of uncertainty quantification. Finally, the framework is extended to a class of generic dams with different heights, concrete strengths, and foundation-to-concrete flexibility. By probabilistic seismic performance evaluation, system safety is guaranteed using the proposed dynamic RBDO compared to the median response in applying a set of as-recorded ground motions. This paper provides a new paradigm for RBDO of structures with dynamic loading and a robust decision-making during dam design.

An Optimized Positioning Algorithm Based on Improved Gaussian Filtering

WiFi offers simple, convenient, ubiquitous, and economic solutions for indoor positioning services, by matching a pre-established WiFi’s RSSI fingerprint database to a mobile terminal’s received RSSI values. A setback of this fingerprint matching method is its low precision, only miserably on order of meters, due to signal impairment by indoor complicated environment. To circumvent this, we revise the traditional weighted K-neighborhood algorithm by incorporating a Bayesian probability optimization. The proposed combination of Bayesian with weighted K-neighborhood algorithm improves the accuracy and reduces the average running time. Computer simulation shows that proposed Bayesian probabilistic optimization algorithm improves the positioning accuracy from 34% to 46%, with an average of about 14.86%, and the computation stability is also enhanced.

Comparison of Feature Selection Techniques for Power Amplifier Behavioral Modeling and Digital Predistortion Linearization.

The power amplifier (PA) is the most critical subsystem in terms of linearity and power efficiency. Digital predistortion (DPD) is commonly used to mitigate nonlinearities while the PA operates at levels close to saturation, where the device presents its highest power efficiency. Since the DPD is generally based on Volterra series models, its number of coefficients is high, producing ill-conditioned and over-fitted estimations. Recently, a plethora of techniques have been independently proposed for reducing their dimensionality. This paper is devoted to presenting a fair benchmark of the most relevant order reduction techniques present in the literature categorized by the following: (i) greedy pursuits, including Orthogonal Matching Pursuit (OMP), Doubly Orthogonal Matching Pursuit (DOMP), Subspace Pursuit (SP) and Random Forest (RF); (ii) regularization techniques, including ridge regression and least absolute shrinkage and selection operator (LASSO); (iii) heuristic local search methods, including hill climbing (HC) and dynamic model sizing (DMS); and (iv) global probabilistic optimization algorithms, including simulated annealing (SA), genetic algorithms (GA) and adaptive Lipschitz optimization (adaLIPO). The comparison is carried out with modeling and linearization performance and in terms of runtime. The results show that greedy pursuits, particularly the DOMP, provide the best trade-off between execution time and linearization robustness against dimensionality reduction.

A probabilistic multimodal optimization algorithm based on Buffon principle and Nyquist sampling theorem for noisy environment

A novel method named probabilistic multimodal optimization (PMO) algorithm is proposed in this paper to competently optimize noisy objective function. In the PMO algorithm, we propose two new strategies to make the algorithm have the capability of probability prediction and multiple extreme points optimization. The first strategy is concerned with the partition strategy of search space based on Buffon principle, which provides the probability prediction of peak detection ratio according to the Buffon distance and extrema resolution. At the same time, several local scopes where multiple extreme points are located can be retained by the partition strategy. The second strategy deals with a same-peak detection method based on Nyquist sampling theorem to identify the position of the extreme point. Based on 12 benchmark functions, experiments are carried out from three aspects, including probabilistic property verification, analysis of the influence of sampling frequency on the same-peak detection method, and multiple extreme points optimization. The validity of PMO algorithm theory and the probabilistic characteristic of global optimization under noise interference are proved, and it is denoted that PMO algorithm can locate more optima and gain higher location precision.

On Random Generation of Supermodular Capacities

Discrete fuzzy measures (capacities) are a powerful tool to model interacting inputs and aggregate them using Choquet, Sugeno, and other fuzzy integrals. Simulation studies and probabilistic optimization algorithms require efficient random sampling from the set of fuzzy measures. It is a challenging problem due to an extremely high complexity of the polytope of capacities. This article addresses random sampling from two special classes: supermodular and submodular fuzzy measures. A suitable marginal contributions representation converts the problem to sampling from an order polytope, which is dealt efficiently using the Markov chain random walk.

Energy-Efficient Analog Combiner Design Using Low-Resolution Phase Shifters and Antenna Selection for mmWave D2D Communications

This study investigates the use of an analog combiner comprising low -resolution phase shifters (PSs) to suppress interference from other devices and improve the received signal power of interest in millimeter-wave device-to-device networks simultaneously. Unlike typical analog combining architectures with PS numbers equaling the number of receiver antennas, the number of PSs in this study is less than that of receiver antennas; this feature saves on equipment costs and improves power efficiency. We jointly optimize the receiver antenna selection and analog combiner design with low-resolution PSs to maximize the spectral efficiency of the resultant device. Because the formulated joint optimization problem is non-convex, we propose a probabilistic evolutionary optimization algorithm with an explicit expression for learning the tilting parameters of the designed sampling distribution. Simulation results demonstrate the excellent performance of the proposed algorithm compared with the semidefinite relaxation approach in terms of spectral and energy efficiency.

Optimal design of Kaibel dividing wall columns based on improved particle swarm optimization methods

Dividing wall columns (DWCs) are able to reduce operating costs and save capital costs for distillation columns, which verifies that DWC is a useful strategy in terms of distillation process intensification. DWCs are better choices than the corresponding conventional distillation sequences from both economical and environmental point of views. As DWC saves energy considerably and reduces CO2 emissions, it proves to be a breakthrough towards sustainable distilling. The four-product Kaibel DWC can further intensify the distillation process, and save energy cost about 40%. However, due to the complicated structures and strong interactions, the optimal design of the four-product Kaibel DWC is challenging to solve through conventional optimization methods. Therefore, this paper investigates the applicability of probabilistic global optimization algorithms, including standard particle swarm optimization (PSO), chaotic PSO, cellular PSO, and quantum-behaved PSO algorithms. This paper aims to explore the most appropriate improved PSO algorithm for the Kaibel DWC. The standard PSO and chaotic PSO show the premature convergence in the optimization process, while the cellular PSO and quantum-behaved PSO can prevent premature convergence. The chaotic PSO algorithm is more appropriate for the optimization of the continuous function, and it is less suitable for the optimization of the Kaibel DWC. The quantum-behaved PSO algorithm is relatively not appropriate for the optimization of the Kaibel DWC because of the long calculation time for rigorous simulation. The cellular PSO algorithm, based on the concept of cellular neighborhood, is the most appropriate algorithm for the optimization of the Kaibel DWC.

Probabilistic optimization algorithms for real-coded problems and its application in Latin hypercube problem

This paper proposes a novel optimization algorithm for read-coded problems called the Probabilistic Optimization Algorithm (POA). In the proposed algorithm, rather than a binary or integer, a probabilistic representation is used for the individuals. Each individual in the proposed algorithm is a probability density function and is capable of representing the entire search space simultaneously. In the search process, each solution performs as a local search and climbs the local optima, and at the same time, the interaction among the probabilistic individuals in the population offers a global search. The parameters of the proposed algorithm are studied in this paper and their effect on the search process is presented. A structured population is proposed for the algorithm and the effect of different structures is analyzed. The algorithm is used to solve Latin Hyper-cube problem and experimental studies suggest promising results. Different benchmark functions are also used to test the algorithm and results are presented. The analyses suggest that the improvement is more significant for large scale problems.

Comparison of the Probabilistic Ant Colony Optimization Algorithm and Some Iteration Method in Application for Solving the Inverse Problem on Model With the Caputo Type Fractional Derivative.

This paper presents the algorithms for solving the inverse problems on models with the fractional derivative. The presented algorithm is based on the Real Ant Colony Optimization algorithm. In this paper, the examples of the algorithm application for the inverse heat conduction problem on the model with the fractional derivative of the Caputo type is also presented. Based on those examples, the authors are comparing the proposed algorithm with the iteration method presented in the paper: Zhang, Z. An undetermined coefficient problem for a fractional diffusion equation. Inverse Probl. 2016, 32.

Calibration of two rock-frame models using deterministic and probabilistic approaches: Application to a carbonate reservoir in south-west Iran

Calibration of the rock-frame elastic properties is an important step in establishing the rock-physics models for quantitative seismic reservoir characterization. While several empirical and theoretical rock-frame models have been proposed in the literature, their applicability to carbonate rocks is challenged due to their specific textural and varied pore type characteristics. Here, the suitability of an empirical model (Nur’s critical porosity), and a semi-empirical inclusion-based model (Keys-Xu) for a hard rock carbonate reservoir in south-west Iran is examined. Wireline log data in conjunction with deterministic and probabilistic optimization algorithms are employed to calibrate the free parameters in these models, as well as the minerals’ effective elastic moduli and density. The optimum values derived from the very fast simulated annealing algorithm were comparable to the estimated values from a probabilistic calibration workflow (Markov chain Monte Carlo sampling-based method). The uncertainty quantification via the probabilistic approach indicates higher uncertainty is associated with the calibrated clay elastic parameters (10%–13%) compared to the uncertainty associated with the calcite elastic moduli (1%–3%) in both rock-frame models. This could be attributed to the dominance of the calcite mineral and lack of variability in clay volume fraction in the interval of interest. The uncertainty associated with critical porosity and pore aspect ratio are 3% and 7%, respectively. The results of this case study show that simple rock frame models with only one free parameter (critical porosity or pore aspect ratio) could successfully model the elastic logs without resorting to more sophisticated rock-frame models.

From BIM to Scan Planning and Optimization for Construction Control

Scan planning of buildings under construction is a key issue for an efficient assessment of work progress. This work presents an automatic method aimed to determinate the optimal scan positions and the optimal route based on the use of Building Information Models (BIM) and considering data completeness as stopping criteria. The method is considered for a Terrestrial Laser Scanner mounted on a mobile robot following a stop & go procedure. The method starts by extracting floor plans from the BIM model according to the planned construction status, and including geometry and semantics of the building elements considered for construction control. The navigable space is defined from a binary map considering a security distance to building elements. After a grid-based and a triangulation-based distribution are implemented for generating scan position candidates, a visibility analysis is carried out to determine the optimal number and position of scans. The optimal route to visit all scan positions is addressed by using a probabilistic ant colony optimization algorithm. The method has been tested in simulated and real buildings under very dissimilar conditions and structural construction elements. The two approaches for generating scan position candidates are evaluated and results show the triangulation-based distribution as the more efficient approach in terms of processing and acquisition time, especially for large-scale buildings.

Phoenics: A Bayesian Optimizer for Chemistry.

We report Phoenics, a probabilistic global optimization algorithm identifying the set of conditions of an experimental or computational procedure which satisfies desired targets. Phoenics combines ideas from Bayesian optimization with concepts from Bayesian kernel density estimation. As such, Phoenics allows to tackle typical optimization problems in chemistry for which objective evaluations are limited, due to either budgeted resources or time-consuming evaluations of the conditions, including experimentation or enduring computations. Phoenics proposes new conditions based on all previous observations, avoiding, thus, redundant evaluations to locate the optimal conditions. It enables an efficient parallel search based on intuitive sampling strategies implicitly biasing toward exploration or exploitation of the search space. Our benchmarks indicate that Phoenics is less sensitive to the response surface than already established optimization algorithms. We showcase the applicability of Phoenics on the Oregonator, a complex case-study describing a nonlinear chemical reaction network. Despite the large search space, Phoenics quickly identifies the conditions which yield the desired target dynamic behavior. Overall, we recommend Phoenics for rapid optimization of unknown expensive-to-evaluate objective functions, such as experimentation or long-lasting computations.