Adaptive Optimization Framework Research Articles

A new adaptive Kriging-based optimization (AKBO) framework for constrained optimization problems: A case study on shared autonomous electric vehicle system design

Adaptive Kriging-based optimization (AKBO) efficiently performs optimization by employing adaptive sampling methods that sequentially update Kriging model. To solve constrained optimization problems, existing adaptive sampling methods add samples based on probability of feasibility (PoF) representing the probability that all constraints are satisfied. However, these methods have the limitation that samples are added inefficiently when the PoF is inaccurate. Therefore, this study proposes a new AKBO framework that is robust to instability of PoF in adaptive sampling. To this end, a near constraint boundary region that is expected to be close to the constraint boundary is defined based on a probability measure. Then, a near constraint boundary search (NCBS) algorithm, which explores the sample point that minimizes the objective function within the near constraint boundary region, is developed for global search, whereas a near optimum search (NOS) algorithm, which searches for better solutions that may exist near the current optimum candidate, is developed for local search. To alleviate the distortions of the Kriging model that may occur due to the highly nonlinear response, a truncated constraint function (TCF) method that limits the extremely violated constraint values used in the design of experiment (DoE) of the Kriging model is also proposed. This study verifies the performance of the proposed AKBO framework through the shared autonomous electric vehicle (SAEV) system design optimization problem. Comparison of optimum designs derived from various optimization methods shows that the proposed AKBO framework derives the feasible global optimum with high accuracy and efficiency.

An adaptive grouping sonar-inertial odometry for underwater navigation

This paper presents a sonar-inertial navigation system with an adaptive grouping optimization framework, which tightly couples inertial measurements with acoustic data from forward-looking sonar (FLS). The system’s front-end uses the sonar-inertial constraint-based feature matching (SICFM) to increase the accuracy and robustness of data association. SICFM is designed by incorporating two novel measures. Firstly, it combines prior motion information to refine the feature search region. Then, a Gaussian-gamma mixture model is formulated to reduce the impact of speckle noise on the selection of matching points. In the back-end, an adaptive grouping factor graph optimization framework is established. A frame manager is designed using the solvability of motion estimation to dynamically group sonar frames, thereby ensuring stable feature tracking for factor graph optimization. Pre-integration-based inertial measurement factor and reprojection-based sonar measurement factor are constructed as constraints in the optimization, which reduces the dependence on prior knowledge of the scene. Furthermore, an elevation angle parameterization is introduced, aiming at the nonlinear projection of sonar features. The evaluation is presented in simulation and real-world experiments, which validate our proposed algorithm in terms of robust localization with high accuracy.

A Rule-Based Approach with Fuzzy Logic for Adaptive Network Optimization

This study introduces an adaptive network optimization framework grounded in fuzzy logic and rule-based machine learning for cognitive radio systems. The primary objective is the concurrent minimization of interference, congestion, and bit error rate, coupled with enhancements in throughput and signal-to-noise ratio. A comprehensive set of rules, guided by linguistic variables for qualitative network aspects, is established alongside membership functions for quantitative analysis.The incorporation of machine learning into our approach enables adaptability to diverse network conditions, resulting in overall improved efficiency. The positive impact of machine learning is particularly evident in the reduction of congestion during specific timeframes. Specifically, on Saturday at hour 6, congestion decreases from the conventional 3.055 to 2.86. This notable improvement underscores the efficacy of machine learning in expediting the sensing mechanism of secondary users, facilitating the rapid identification of unused channels from primary users. The findings contribute to the advancement of cognitive radio systems, providing a robust and adaptable solution to address the intricate dynamics of modern wireless networks.

An adaptive multi-fidelity optimization framework based on co-Kriging surrogate models and stochastic sampling with application to coastal aquifer management

Surrogate modelling has been used successfully to alleviate the computational burden that results from high-fidelity numerical models of seawater intrusion in simulation-optimization routines. Nevertheless, little attention has been given to multi-fidelity modelling methods to address cases where only limited runs with computationally expensive seawater intrusion models are considered affordable imposing a limiting factor for single-fidelity surrogate-based optimization as well. In this work, a new adaptive multi-fidelity optimization framework is proposed based on co-Kriging surrogate models considering two model fidelities of seawater intrusion. The methodology is tailored to the needs of solving pumping optimization problems with computationally expensive constraint functions and utilizes only small high-fidelity training datasets. Results from both hypothetical and real-world optimization problems demonstrate the efficiency and practicality of the proposed framework to provide a steep improvement of the objective function while it outperforms a comprehensive single-fidelity surrogate-based optimization method. The method can also be used to locate optimal solutions in the region of the global optimum when larger high-fidelity training datasets are available.

Adaptive rail transit network operations with a rollout surrogate-approximate dynamic programming approach

This paper presents an adaptive optimization framework for dynamic rail transit network operations with use of a rollout surrogate-approximate dynamic programming method. The optimization algorithm derives coordinated decisions of service schedules and train unit deployment with respect to prevailing passenger demand. Considering the computational effectiveness needed for real-time applications, a state-dependent surrogate function is incorporated to approximate the costs associated with operational decisions over future stages. The surrogate approximation is updated iteratively via a temporal difference learning process with feeding of observations made from the transit network. The proposed framework is implemented and tested on a real-world scenario in Hong Kong Light Rail Transit (LRT) network. The results reveal that the proposed framework is able to reduce significantly the total passengers’ waiting times over existing plans with reasonable computational time via use of the surrogate approximation. This suggests the potential of the proposed optimizer for real time applications in large-scale rail transit networks.

Safe Optimization of an Industrial Refrigeration Process Using an Adaptive and Explorative Framework

Real-time optimization can play a key role for improving the performance of industrial processes, but it becomes challenging in cases where the process characteristics are uncertain, particularly when the uncertain characteristics impact safety-related constraints. In this study, we present an adaptive and explorative real-time optimization framework that can effectively learn the characteristics of an industrial refrigeration plant through externally driven changes in cooling load targets and through self-exploration. We utilize Gaussian processes (GP) to facilitate learning unknown compressor characteristics of the plant and we leverage the uncertainty quantification of the GP to drive exploration using a weighted sum term in the objective function for real-time optimization. Furthermore, the uncertainty information is used to probabilistically enforce the maximum total power consumption constraint for the compressors with high confidence at all times. Our simulated experiments demonstrate that the proposed approach safely enhances the energy efficiency of the refrigeration process, closely approximating the performance of a best-case solution that has complete information about the plant performance characteristics. We also demonstrate the impact of varying the exploration term in the solution and how the uncertainty of plant behaviour is reduced even in the absence of cooling load target changes.

A novel adaptive optimization framework for SVM hyper-parameters tuning in non-stationary environment: A case study on intrusion detection system

Building an Intrusion Detection System (IDS) in non-stationary environment is challenging because, in such an environment, intrusion-related data grow every day. A machine learning model trained in a stationary environment where training data does not change, often fails to retain its performance in real world environment. This is because dynamism in data makes the hyper-parametric values of the underlying classifier shift in the search space. For making such a model work for intrusion detection in non-stationary environment, one must have to run hyper-parametric optimization algorithm again and again at various time instances. But the expansion of the existing data in non-stationary environment, makes such a way of tunning the hyper-parameters computationally expensive. So, there is a requirement of more adaptive and computationally efficient optimization frameworks for hyper-parameters to build IDS in non-stationary environment. This paperwork proposes a novel framework to train a Support Vector Machine (SVM) for intrusion detection by optimizing its hyper-parameters C and γ dynamically. For designing this framework, we have used Moth-Flame Optimization (MFO) as the base optimization algorithm which can be run with random initialization. Further, for utilizing the knowledge generated by running the base optimization algorithm, we have introduced two algorithms- a Lightweight MFO and a simple Knowledge-Based Search. The Lightweight MFO uses the knowledge for initializing the starting solutions and the Knowledge-Based Search uses the knowledge as search space. Based on the result of a drift detection module, the proposed framework identifies the most appropriate algorithm to be used at a particular time instance when re-training of the model is required due to the change in the data. Results have shown a significant reduction in the average time complexity of the hyper-parametric optimization process. We have evaluated our proposed framework on benchmark NSL-KDD dataset and got significantly encouraging convergence rate and detection performance. The obtained average accuracy for IDS built using our proposed framework is 97.5%. Further, we have also compared our framework by considering other metaheuristic algorithms as base optimization algorithms and found that our proposed framework, which uses MFO as a base optimization algorithm outperforms the others.

Online environmentally adaptive trajectory planning for rotorcraft unmanned aerial vehicles

PurposeThis paper aims to propose an environmentally adaptive trajectory planning system considering the dynamic characteristics of unmanned aerial vehicles (UAVs) and the distance between obstacles and the UAV. The system generates a smooth and safe flight trajectory online.Design/methodology/approachFirst, the hybrid A* search method considering the dynamic characteristics of the quadrotor is used to find the collision-free initial trajectory. Then, environmentally adaptive velocity cost is designed for environment-adaptive trajectory optimization using environmental gradient data. The proposed method adaptively adjusts the autonomous flight speed of the UAV. Finally, the initial trajectory is applied to the multi-layered optimization framework to make it smooth and dynamically viable.FindingsThe feasibility of the designed system is validated by online flight experiments, which are in unknown, complex situations.Practical implicationsThe proposed trajectory planning system is integrated into a vision-based quadrotor platform. It is easily implementable onboard and computationally efficient.Originality/valueA hybrid A* path searching method is proposed to generate feasible motion primitives by dispersing the input space uniformly. The proposed method considers the control input of the UAV and the search time as the heuristic cost. Therefore, the proposed method can provide an initial path with the minimum flying time and energy loss that benefits trajectory optimization. The environmentally adaptive velocity cost is proposed to adaptively adjust the flight speed of the UAV using the distance between obstacles and the UAV. Furthermore, a multi-layered environmentally adaptive trajectory optimization framework is proposed to generate a smooth and safe trajectory.

An adaptive context-aware optimization framework for multimedia adaptation service selection

An adaptive context-aware optimization framework for multimedia adaptation service selection

Chance-constrained optimization-based solar microgrid design and dispatch for radial distribution networks

We consider a solar microgrid design and dispatch problem using an adaptive stochastic optimization framework. First, we propose a two-stage mixed-integer model for optimal placement and planning of distributed generation (DG) units and energy storage system units. We incorporate time series modeling into stochastic optimization approach to characterize the solar irradiance uncertainty. In the first stage, design decisions (e.g., location and sizing of DGs) are made and in the second stage, dispatch decisions (e.g., how much to generate, how much to store) are made such that electricity demand is met in a reliable and cost effective way. Chance constraints are employed to control the real load shedding within a predefined probability level. Then, we propose a combined sample average approximation (SAA) and linearization technique to solve this problem more efficiently. The advantage of this approach is that no additional binary variables are introduced while reformulating the chance constraints. Computational time, quality of solution, and load shedding percentage are compared with the traditional SAA. Moreover, we carry out a comprehensive out-of-sample simulation on a real-world case study in the state of Arizona assessing the effectiveness of our approach. Numerical experiments demonstrate that the chance constraints are effective tools for control of load shedding in distributed generation and the proposed approach outperforms traditional methods both in terms of true load shedding percentage and computational time.

Data-driven experimental design and model development using Gaussian process with active learning

Interest in computational modeling of cognition and behavior continues to grow. To be most productive, modelers should be equipped with tools that ensure optimal efficiency in data collection and in the integrity of inference about the phenomenon of interest. Traditionally, models in cognitive science have been parametric, which are particularly susceptible to model misspecification because their strong assumptions (e.g. parameterization, functional form) may introduce unjustified biases in data collection and inference. To address this issue, we propose a data-driven nonparametric framework for model development, one that also includes optimal experimental design as a goal. It combines Gaussian Processes, a stochastic process often used for regression and classification, with active learning, from machine learning, to iteratively fit the model and use it to optimize the design selection throughout the experiment. The approach, dubbed Gaussian process with active learning (GPAL), is an extension of the parametric, adaptive design optimization (ADO) framework (Cavagnaro, Myung, Pitt, & Kujala, 2010). We demonstrate the application and features of GPAL in a delay discounting task and compare its performance to ADO in two experiments. The results show that GPAL is a viable modeling framework that is noteworthy for its high sensitivity to individual differences, identifying novel patterns in the data that were missed by the model-constrained ADO. This investigation represents a first step towards the development of a data-driven cognitive modeling framework that serves as a middle ground between raw data, which can be difficult to interpret, and parametric models, which rely on strong assumptions.

Adaptive Framework for Structural Pattern Optimization

This research outlines a new computational workflow for the design and optimization of patterned, perforated surface structures. Well-designed surface structures can be highly efficient on their own, but their potential for structural efficiency can be notably improved by deliberately introducing specific aperture patterns. Patterned surfaces can also be used to produce more stimulating architectural environments, and even those of increasing complexity can be realized thanks to recent developments in digital fabrication. With this said, designers currently lack a streamlined and rigorous approach for the exploration of patterned surface structures. This research aims to address this issue by advancing a recent work that employs NURBS-based isogeometric analysis to integrate structural analysis into an accessible CAD modeling platform. Specifically, this paper proposes an adaptive pattern optimization framework formulated to save designers appreciable computational time. Not only does this framework offer a way to quickly visualize various design solutions, but it also provides the designer an important opportunity for design interaction and control over design evolution, in turn lending itself as a versatile tool for the exploration and conceptual design of patterned surface structures.

A framework for adaptive open-pit mining planning under geological uncertainty

Mine planning optimization aims at maximizing the profit obtained from extracting valuable ore. Beyond its theoretical complexity—the open-pit mining problem with capacity constraints reduces to a knapsack problem with precedence constraints, which is NP-hard—practical instances of the problem usually involve a large to very large number of decision variables, typically of the order of millions for large mines. Additionally, any comprehensive approach to mine planning ought to consider the underlying geostatistical uncertainty as only limited information obtained from drill hole samples of the mineral is initially available. In this regard, as blocks are extracted sequentially, information about the ore grades of blocks yet to be extracted changes based on the blocks that have already been mined. Thus, the problem lies in the class of multi-period large scale stochastic optimization problems with decision-dependent information uncertainty. Such problems are exceedingly hard to solve, so approximations are required. This paper presents an adaptive optimization scheme for multi-period production scheduling in open-pit mining under geological uncertainty that allows us to solve practical instances of the problem. Our approach is based on a rolling-horizon adaptive optimization framework that learns from new information that becomes available as blocks are mined. By considering the evolution of geostatistical uncertainty, the proposed optimization framework produces an operational policy that reduces the risk of the production schedule. Our numerical tests with mines of moderate sizes show that our rolling horizon adaptive policy gives consistently better results than a non-adaptive stochastic optimization formulation, for a range of realistic problem instances.

Impacts of forest tax under timber price uncertainty

We examine the effect of forest taxes on the optimal harvest decision and the forest land value using an adaptive optimization framework– the stand is harvested when the stumpage price is equal or higher than an age-dependent reservation price. We prove that the site value tax or site productivity tax does not change the optimal reservation prices; the timber value tax reduces the optimal reservation prices; the yield tax increases the optimal reservation prices when the non-timber benefits are increasing with stand age, or when they are independent of the stand age and the regeneration cost is positive; and the unit harvest tax increases the reservation prices when stand age is relatively high and could reduce the reservation price when stand age is low. Numerical results show that the optimal harvest age is only sensitive to the unit and timber value tax. The tax burden is reduced for all types of taxes except the site value tax. Under a taxation budget constraint, recognition of stumpage price uncertainty may change forest owners’ preferences for different types of taxes.

Adaptive Optimization Framework for Control of Multi-Agent Systems

The main focus of this work is an optimization-based framework for control of multi-agent systems that synthesizes actions steering a given system towards a specified state. The primary motivation for the research presented is a fascination with birds, which save energy on long-distance flights via forming a V-shape. We ask the following question: Are V-formations a result of solving an optimization problem and can this concept be utilized in multi-agent systems, particularly in drones swarms, to increase their safety and resilience? We demonstrate that our framework can be applied to any system modeled as a controllable Markov decision process with a cost (reward) function. A key feature of the procedure we propose is its automatic adaptation to the performance of optimization towards a given global objective. Combining model-predictive control and ideas from sequential Monte-Carlo methods, we introduce a performance-based adaptive horizon and implicitly build a Lyapunov function that guarantees convergence. We use statistical model-checking to verify the algorithm and assess its reliability.

Adaptive Network Resource Optimization for Heterogeneous VLC/RF Wireless Networks

Deploying a radio frequency (RF) access point (AP) to the visible light communication (VLC) system is a promising strategy to overcome the VLC’s limitations, such as limited coverage, strictly line-of-sight transmission, and mobility robustness, etc. In this paper, we focus on the energy-aware design of network selection and resource allocation for a heterogeneous network combining with RF and VLC APs. For adapting to different timescale network states and stochastic data arrival, we propose an on-line two-timescale adaptive network resource optimization (ANRO) framework by employing the Lyapunov optimization technique. At the large timescale, we first develop a closed-form solution for the subproblem of network selection for user equipment. Second, we design a cost-effective and easy-to-realize algorithm for VLC’s joint transmission scheduling and power control subproblem, which is a nonconvex optimization. While at the small timescale, we obtain the optimal solution for RF’s joint resource block and power allocation subproblem, which is proven a mixed integer nonlinear optimization. Simulation results demonstrate that the ANRO can achieve a tradeoff between network power consumption and delay. Furthermore, it not only can stabilize the network but also can significantly reduce the energy consumption compared with other existing schemes.

Robust aerodynamic shape design based on an adaptive stochastic optimization framework

Optimization techniques combined with uncertainty quantification are computationally expensive for robust aerodynamic optimization due to expensive CFD costs. Surrogate model technology can be used to improve the efficiency of robust optimization. In this paper, non-intrusive polynomial chaos method and Kriging model are used to construct a surrogate model that associate stochastic aerodynamic statistics with airfoil shapes. Then, global search algorithm is used to optimize the model to obtain optimal airfoil fast. However, optimization results always depend on the approximation accuracy of the surrogate model. Actually, it is difficult to achieve a high accuracy of the model in the whole design space. Therefore, we introduce the idea of adaptive strategy to robust aerodynamic optimization and propose an adaptive stochastic optimization framework. The surrogate model is updated adaptively by increasing training airfoils according to historical optimization results to guarantee the accuracy near the optimal design point, which can greatly reduce the number of training airfoils. The proposed method is applied to a robust aerodynamic shape optimization for drag minimization considering uncertainty of Mach number in transonic region. It can be concluded that the proposed method can obtain better optimal results more efficiently than the traditional robust optimization method and global surrogate model method.

Energy-Efficient Images.

In this paper, a novel method is presented for producing energy-efficient images, i.e., images that consume less electrical energy on energy-adaptive displays, yet have the same or very similar perceptual quality to their original images. The proposed method relies on the fact that the energy consumption of pixels in modern energy-adaptive displays like OLED displays is directly proportional to the luminance of the pixels. Hence, in this paper, to reduce the energy consumption of an image, while at the same time preserving its perceptual quality, it is proposed to reduce the luminance of the pixels in the image by one just-noticeable-difference (JND) threshold. To determine the JND thresholds, an adaptive saliency-modulated JND (SJND) model is developed. In the proposed model, the JND thresholds of each block in the given image are elevated by two non-linear saliency modulation functions using the visual saliency of the block. The parameters of the saliency modulation functions are estimated through an adaptive optimization framework, which utilizes a state-of-the-art saliency-based objective image quality assessment method. To evaluate the proposed methods, a set of subjective experiments were conducted, and the real energy consumption of the produced energy-efficient images were measured by an accurate power monitor equipment on an OLED display. The obtained experimental results demonstrated that, on average, the proposed method is able to reduce the energy consumption by about 14.1% while preserving the perceptual quality of the displayed images.

AMOF: adaptive multi-objective optimization framework for coverage and topology control in heterogeneous wireless sensor networks

In this paper, we propose adaptive multi-objective optimization framework based on non-dominated sorting genetic algorithm-II and learning automata (LA) for coverage and topology control in heterogeneous wireless sensor networks. The multi-objective optimization approach of the proposed framework, called MOOCTC (multi-objective optimization coverage and topology control), can simultaneously optimize several conflicting issues such as number of active sensor nodes, coverage rate of the monitoring area and balanced energy consumption while maintaining the network connectivity. This approach incorporates problem-specific knowledge in its operators to find high-quality solutions. In addition, this approach uses LA to dynamically adapt the crossover and mutation rates without any external control to improve the behavior of the optimization algorithm. Simulation results demonstrate the efficiency of the proposed multi-objective optimization approach in terms of lifetime, coverage and connectivity.

Simulation-based adaptive calibration and optimization of intelligent transportation systems for highway congestion management

The paper introduces an adaptive optimization framework that is used to calibrate and optimize multiple intelligent transportation strategies such as high occupancy toll (HOT), ramp metering, and variable speed limits. The framework is capable of estimating the OD (origin–destination) matrix, calibrating the model with the experimental data, and conducting system optimization in transportation. This capability is shown with simulation experiments based on the public data of I-95 Florida. The preliminary results show that the revenue increases with target speed up to a point, after which the revenue drops. Also, the throughput on the HOT increases with decreasing target speed. So, to maximize revenue and throughput together, the optimal values are observed at moderate levels of the feasible range. Also, in the presence of accidents, the total throughput is improved by decreasing the toll rate. Optimal level of vehicles on the HOT lanes is examined to minimize vehicle emissions.