Heuristic Optimization Method Research Articles

A parallel variable population multi-objective optimizer for accelerator beam dynamics optimization

The simultaneous optimization of multiple objective functions is needed in many particle accelerator applications. In this paper, we present a parallel evolution based multi-objective optimizer that uses a variable population from generation to generation and an external storage to save good solutions. Two heuristic optimization methods, one uses the unified differential evolution and the other uses the real-coded genetic algorithm, are included in the optimizer to generate next generation candidate solutions, and are compared in the test examples. As an application, we applied this optimizer to the beam dynamics design optimization of a photoinjector and attained the optimal front solutions after 200 generations with the unified differential evolution offspring production scheme.

Benchmarking of Heuristic Algorithms for Energy Router-Based Packetized Energy Management in Smart Homes

This article presents an ER-based PEM strategy for PV integrated smart homes to jointly optimize their load scheduling delays, energy transactions cost, and battery degradation cost. The proposed approach incorporates a MA case, where, the ER acts as a main selecting agent realized by all other system elements. This leads to a combinatorial optimization problem, which can be effectively solved by heuristic optimization methods (HOMs), namely, genetic algorithm (GA), binary particle swarm optimization (BPSO), differential evolution (DE) algorithm, and harmony search algorithm (HSA). Specifically, we investigate the impact of the hyperparameters of the HOMs on the designed ER-based PEM system. Simulations are carried out for multiple smart homes under varying weather conditions to evaluate the effectiveness of HOMs in terms of selected performance metrics. Results show that the ER-based PEM reduces the average aggregated system cost, ensures economic benefits by selling surplus energy, while meeting customers energy packet demand, satisfying their quality-of-service, and operational constraints.

Parameter Space Exploration of Cellular Mechanical Metamaterials Using Genetic Algorithms

Cellular materials widely exist in natural biologic systems such as honeycombs, bones, and woods. With advances in additive manufacturing, research on cellular metamaterials is emerging due to their unique mechanical performance. However, the design of on-demand cellular metamaterials usually requires solving a challenging inverse design problem for exploring complex structure–property relations of microstructured representative volume elements (RVEs) in the design domain. Here, we propose an experience-free and systematic methodology for exploring a parametrized system for microstructures of cellular mechanical metamaterials using a multiobjective genetic algorithm (GA). Globally, by considering the importance of the initial population selection for a population-based heuristic optimization method, we study the impact of the populations initialized by the different sampling methods on the optimal solutions. Locally, we develop our method by using a micro-GA with a new searching strategy, which requires the standard genetic algorithm to be conditionally run for a sufficient number of times with a small population size during the global searching process. We have applied our method to explore optimal solutions for applications mapped on two different parameter spaces of the cellular mechanical metamaterials with periodic and nonperiodic RVEs effectively and accurately.

Localization and Sizing of Distributed Generation through a Genetic Algorithm to Improve Voltage Profile Using Ecuadorian Standards

The approach proposed for the development of this research work was based on the integration of Distributed Energy Generation (DG) into an energy distribution network, improving the voltage profile by establishing an optimal location and sizing of DG power plants, for which the use of a heuristic optimization method known as Genetic Algorithm (GA), which has several restrictions to limit its operation and thus achieve an optimal solution to the planned optimization problem, was resorted to. The development of a power flow allowed having the voltage measurement in the bars of the network before incorporating DG, allowing to verify that the voltage fluctuations present an improvement after the incorporation one by one of each DG central defined by the GA. The systems in which the voltage measurement was performed, as well as the verification of its fluctuations with and without DG once the GA was used, were the IEEE systems of 9 and 14 busbars, the latter the one used to demonstrate the scalability of the GA to locate and dimension DG without considering the number of busbars in a system.

Parameter Estimation of The Blumberg Model Using Simulated Annealing Algorithm: Case Study of Broiler Body Weight

The Blumberg model is one of the logistic models. The advantage of the Blumberg model is the flexibility of the inflection point. The Blumberg model is believed to be suitable for modeling the growth of living organs. In this article, we estimate the parameters of the Blumberg model using simulated annealing algorithm. The simulated annealing algorithm is a heuristic optimization method based on the metal annealing process. The data used is Broiler daily weight data. The model obtained fits the daily weight data of Broiler. Our results show that the closer the cooling schedule factor to 1, the smaller the error. In addition, we must carefully select the initial temperature. The selection of the initial temperature that is not suitable drives the error to enlarge.

Optimizing Optical Potentials With Physics-Inspired Learning Algorithms

We present our experimental and theoretical framework, which combines a broadband superluminescent diode with fast learning algorithms to provide speed and accuracy improvements for the optimization of on-dimensional optical dipole potentials, here generated with a digital micromirror device. To characterize the setup and potential speckle patterns arising from coherence, we compare the superluminescent diode to a single-mode laser by investigating interference properties. We employ machine-learning tools to train a physics-inspired model acting as a digital twin of the optical system predicting the behavior of the optical apparatus including all its imperfections. Implementing an iterative algorithm based on iterative learning control we optimize optical potentials an order of magnitude faster than heuristic optimization methods. We compare iterative model-based ``offline'' optimization and experimental feedback-based ``online'' optimization. Our methods provide a route to fast optimization of optical potentials, which is relevant for the dynamical manipulation of ultracold gases.

Improving structural health monitoring with impact zone localization using heuristic algorithms

AbstractThis paper presents a novel approach for reconstructing the characteristics of non-punctual impact events on elastic plates by introducing multi-parameter optimization. The objective function is minimized using two heuristic optimization techniques, particle swarm optimization, and differential evolution, to reconstruct impact force characteristics. The force was regarded as taking the form of a uniform pressure over a part of the plate called the patch. The Maxwell-Betti theorem was considered to decouple the problem of localization and the time history of the applying unknown load. The approach based on heuristic optimization methods has been proven a performance to locate the impact zone. A comparison between particle swarm optimization and differential evolution was discussed.

Estimation of third body efficiencies from experimental data: Application to hydrogen combustion

In this work, the role of H2O and CO2 as diluents is investigated, in operating conditions relevant to MILD combustion of hydrogen. Through virtual species analysis, the role of third body collisions emerges in phenomena such as ignition and speciation. Global sensitivity analysis shows that a single third body efficiency exhibits a larger impact than a group of sensitive reactions. For fall-off reactions, the adoption of a PLOG format, rather than a classical TROE can introduce errors as high as 2000% on macroscopic combustion targets due to the absence of key collision efficiencies. Heuristic optimization methods in OptiSMOKE++ provide best estimates and confidence regions for the third body efficiencies of H2O and CO2 in H+O2(+M)⇋HO2(+M) and H2O2(+M)⇋2OH(+M) in TROE formulation. Eventually, a method to extract this information using the PLOG formulation is proposed. This approach is applicable to other reactions in case theoretical calculations are unavailable or difficult to perform.

MR-DRO: A Fast and Efficient Task Offloading Algorithm in Heterogeneous Edge/Cloud Computing Environments

With the rapid development of Internet of Things (IoT) and next-generation communication technologies, resource-constrained mobile devices (MDs) fail to meet the demand of resource-hungry and compute-intensive applications. To cope with this challenge, with the assistance of mobile-edge computing (MEC), offloading complex tasks from MDs to edge cloud servers (CSs) or central CSs can reduce the computational burden of devices and improve the efficiency of task processing. However, it is difficult to obtain optimal offloading decisions by conventional heuristic optimization methods, because the decision-making problem is usually NP-hard. In addition, there are shortcomings in using intelligent decision-making methods, e.g., lack of training samples and poor ability of migration under different MEC environments. To this end, we propose a novel offloading algorithm named meta reinforcement-deep reinforcement learning-based offloading, consisting of a meta-reinforcement learning (meta-RL) model, which improves the migration ability of the whole model, and a deep reinforcement learning (DRL) model, which combines multiple parallel deep neural networks (DNNs) to learn from historical task offloading scenarios. Simulation results demonstrate that our approach can effectively and efficiently generate near-optimal offloading decisions in IoT environments with edge and cloud collaboration, which further improves the computational performance and has strong portability when making offloading decisions.

Integration of machine learning prediction and heuristic optimization for mask delivery in COVID-19.

The novel coronavirus pneumonia (COVID-19) has created huge demands for medical masks that need to be delivered to a lot of demand points to protect citizens. The efficiency of delivery is critical to the prevention and control of the epidemic. However, the huge demands for masks and massive number of demand points scattered make the problem highly complex. Moreover, the actual demands are often obtained late, and hence the time duration for solution calculation and mask delivery is often very limited. Based on our practical experience of medical mask delivery in response to COVID-19 in China, we present a hybrid machine learning and heuristic optimization method, which uses a deep learning model to predict the demand of each region, schedules first-echelon vehicles to pre-distribute the predicted number of masks from depot(s) to regional facilities in advance, reassigns demand points among different regions to balance the deviations of predicted demands from actual demands, and finally routes second-echelon vehicles to efficiently deliver masks to the demand points in each region. For the subproblems of demand point reassignment and two-batch routing whose complexities are significantly lower, we propose variable neighborhood tabu search heuristics to efficiently solve them. Application of the proposed method in emergency mask delivery in three megacities in China during the peak of COVID-19 demonstrated its significant performance advantages over other methods without pre-distribution or reassignment. We also discuss key success factors and lessons learned to facilitate the extension of our method to a wider range of problems.

Research on simplification of branches method of accident sequences based on expert knowledge and heuristic optimization algorithm

The simplification of branches method of accident sequences based on expert knowledge and heuristic optimization algorithm is proposed in this paper, to increase the effectiveness of nuclear power plant safety margin analysis. First, the classification criteria for accident sequence and expert knowledge are used to screen out the deterministic failure sequences and deterministic safety sequences that are obviously. Secondly, a heuristic optimization method is used to study the variation in the accident sequences for those accident sequences that are challenging to evaluate with expert knowledge. Moreover, a hybrid particle swarm optimization (HPSO) is proposed to increase the accuracy of key safety parameters. This enables efficient accident sequence type identification, by comparing and evaluating key safety parameters and safety limits. Finally, the accident sequences are simplified using the cut-off frequency. To verify the effectiveness of the method, the small break loss of coolant (SBLOCA) accident in the nuclear power plant is analyzed in this paper. The numerical results show that the proposed HPSO improved the accuracy of solving the key safety parameters of accident sequences compared with the traditional method. By simplifying the accident sequences, the number of accident sequences requiring further analysis was reduced from 64 to 5.

Composite Lamina Model Design with the Use of Heuristic Optimization.

In engineering practice, a problem is quite often faced in which the number of unknown parameters exceeds the number of conditions or requirements or, otherwise, there are too many requirements for too few parameters to design. Such under- or over-defined tasks are sometimes not possible to solve using a direct approach. The number of solutions to such problems is multiple, and it is most rational to search for the optimal one by numerical methods since the more unknown design parameters there are to be designed, the more potential solutions there are. This article discusses a way to find an optimal solution to such an underdetermined problem by heuristic optimization methods on the basis of the example of designing a composite wing skin of an aircraft. Several heuristic approaches, specifically gradient descent and Tabu search, are studied to solve the design problem and to locate an optimal solution. They are also compared to a conventional direct approach. The examined composite lamina is optimized by the target function of minimum weight with the constraints of strength and buckling failure criteria. In most of the observed cases, the heuristic method designed structures which were considerably better than the structures that were obtained by conventional direct approaches in terms of the weight to load ratio.

Fuzzy optimization of route networks urban passenger transport

The work aims to study methods for optimizing the route networks of urban passenger transport with fuzzy initial information since the process of passenger transportation in cities has many factors of uncertainty and stochasticity and is quite complex, including a whole range of initial data. Heuristic fuzzy optimization methods are proposed. The study's main result is the improvement of methods for optimizing the route networks of urban passenger transport. These methods today make it possible to take into account the opposing interests of the participants in the process of transporting passengers in the city, as well as to set a large set of initial parameters and restrictions for the mathematical model of optimizing the urban route network. A fuzzy transport problem has been solved, a computational experiment has been carried out, and results have been obtained to illustrate the proposed method. This method is easy to understand and apply to find the optimal solution in fuzzy transport models of real problems. An algorithm for finding L alternate routes and determining the ideal route for vehicle traffic control based on a graph database allows vehicle traffic control.

Joint Transmissive and Reflective RIS-Aided Secure MIMO Systems Design Under Spatially-Correlated Angular Uncertainty and Coupled PSEs

This paper investigates a joint transmissive and reflective reconfigurable intelligent surfaces (RIS) -aided secure multiple-input multiple-output (MIMO) system, where both a RIS-assisted transmitter and a RIS-based reflector are deployed to defend against the simultaneous jamming attack and wiretapping threat. Our design focuses on maximizing the sum rate under the unknown jammer’s beamforming, joint RISs’ coupled phase shift errors (PSEs), and spatially-correlated angular channel uncertainties. Besides, we take into account the various quality-of-service (QoS) requirement constraints for guaranteeing the secure performance. Since the problem is non-convex and mathematically intractable, a new optimization framework is established to facilitate the solution development to the formulated problem. Specifically, armed with the Akaike information criterion, a novel diagonalization method is first proposed to estimate the unknown jamming covariance matrix. Then, a series of fractional-eliminated rate expressions is derived that facilitates the application of the proposed Double Deterministic Transformation (DDT) to tackle the coupled stochastic PSEs. Besides, regardless of the spatial correlation matrix, a general discretization method is proposed to convert the e spatially-correlatd angular uncertainties into a worst-case robust one. Subsequently, building upon the above transformations which transform the original problem into tractable one, a two-layer iterative Lagrange multiplier algorithm capitalizing a low-complexity dual method is proposed to obtain the globally optimal solution of the digital precoder, where the multiple QoS constraints are handled without iteration. Meanwhile, we develop a novel polyblock-based multiple penalty method to obtain the globally optimal solutions to RISs’ phase shifts which can simultaneously satisfy the multiple QoS constraints. Moreover, to address the narrow feasibility region induced by the multiple QoS constraints, a heuristic initial optimization method is proposed, which strengthens the existing result. Finally, theoretical analysis and numerical results demonstrate the optimality and the excellent performance of our proposed optimization framework.

THE SHUFFLE FROG LEAPING ALGORITHM FOR THE PRODUCTION LOCATION PROBLEM

The production location problem is one of the most well-known discrete optimization problems. There are many options for setting this task. As a rule, all these variants of the problem of location of production belong to the class of NP-hard problems, that is, for its exact solution of such a problem, algorithms of polynomial complexity are currently unknown. So far, no effective methods have been developed for calculating the lower bounds for this problem, which make it possible to assess the achievement of the optimum. Exact algorithms for this problem are reduced to a complete enumeration of options. In this regard, the use of exact algorithms for solving the problem of production location often turns out to be inappropriate and impossible due to the large time costs. Therefore, the development and study of heuristic optimization methods is of considerable interest. One of the promising areas is the development of algorithms based on well-known metaheuristic approaches that are successfully used to solve many discrete optimization problems. In this paper, we show that one of the classes of production location problems is reduced to the problem of covering a complete graph with vertex-disjoint stars. It is proved that in this formulation the problem can be considered as an optimization problem on an oriented fragmentary structure. This allows you to create hybrid algorithms for finding suboptimal solutions to problems of this class based on a combination of well-known metaheuristics and a fragmentary algorithm. The mixed jumping frogs algorithm was chosen as a metaheuristic. It is shown that this algorithm can be used to find suboptimal solutions on a set of permutations. On the other hand, in the presence of an oriented fragmentary structure, the discrete optimization problem can be reduced to an optimization problem on a set of permutations. Thus, a simple and rather effective method for finding suboptimal solutions to the production location problem has been obtained. The method can be easily transferred to other classes of discrete optimization problems, which can be considered as problems on an oriented fragmentary structure.

Design, Fabrication, and Measurement of Efficient Beam-Shaping Reflectors for 5G Applications

An optimization procedure to design passive metallic reflectors for the fifth-generation (5G) mobile network applications, as well as fabrication and measurement of the designed reflectors, are presented. These reflecting surfaces can be used as passive repeaters in both indoor and outdoor applications to change field coverage by redirecting incident beams into desired directions with relatively controllable power distributions. The design procedure involves an efficient combination of a heuristic optimization method [based on genetic algorithms (GAs)] and a full-wave electromagnetic solver [based on the multilevel fast multipole algorithm (MLFMA)]. The optimized geometries are realized using a 3-D printing method that enables low-cost and adaptive fabrication. During optimizations, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$10\lambda $ </tex-math></inline-formula> by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$10\lambda $ </tex-math></inline-formula> compact and efficient reflectors are implemented as reference designs, and they are fabricated to have 18 GHz center frequency of operation. Measurements in a free-space setup are in good agreement with simulations, demonstrating the success of both the design procedure and the designed reflectors as potential 5G components.

Feasibility study of a practical causal rate-independent damping device for the improved performance of seismic isolated structures

When rate-independent linear damping (RILD) is incorporated into a base-isolated structure, it achieves a similar response displacement reduction effect with significantly lower floor response acceleration compared with linear viscous damping (LVD) with the same loss factor. To address the undesirable stiffness added to an isolation layer when we adopt a mechanical device comprising a few branches of a spring–dashpot link to realize RILD, this study proposes canceling the undesirable isolator stiffness by adding an inerter element and reducing the isolator stiffness. A heuristic optimization method was developed to design the proposed mechanical device to mimic the behavior of RILD. A parametric survey on the optimal designs of the device suggests that the three branches of the spring–dashpot link work best in terms of practicality and feasibility. Seismic control performance was assessed using a 10-story base-isolated building mounted on linear and nonlinear isolation systems. Analyses of linear systems revealed that the proposed device achieved lower inter-story drifts and approximately 40% reduction in floor response acceleration with a 4% increase in isolator displacement compared with the LVD when subjected to high-frequency ground motions. Analyses of the nonlinear systems showed that the displacement control performance of the proposed model was slightly compromised when subjected to ground motions dominated by low-frequency components, thereby identifying a further challenge in developing a causal RILD device for nonlinear structures.

Wireless Sensor Network Coverage Optimization: Comparison of Local Search-Based Heuristics

The Maximum Lifetime Coverage Problem (MLCP) requires heuristic optimization methods due to its complexity. A real-world problem model determines how a solution is represented and the operators applied in these heuristics. Our paper describes adapting a local search scheme and its operators to MLCP optimization. The operators originate from three local search algorithms we proposed earlier: LSHMA, LSCAIA, and LSRFTA. Two steps of the LS scheme’s main loop can be executed in three different ways each. Hence, nine versions of the LS approach can be obtained. In experimental research, we verified their effectiveness. Test cases come from three benchmarks: SCP1, proposed and used in our earlier research on the three LS algorithms mentioned above, and two others found in the literature. The results obtained with SCP1 showed that the algorithm based on the hypergraph model approach (HMA) is the most effective. The remaining results of the other algorithms divide them into two groups: effective ones and weak ones. However, other benchmarks showed that the more redundant the coverage of points of interest (POIs) by sensors, the more effective the perturbation method from the approach inspired by cellular automata (CAIA). The findings expose the strengths and weaknesses of the problem-specific steps applied in the LS algorithms.

Genetic algorithm-based adaptive weighted fuzzy logic control (awFLC) for traction power control

One of the most important issues for FLC systems is the problem of finding the right balance between interpretability and accuracy. For this delicate balance, several methods which can be integrated into fuzzy logic, and tune the fuzzy logic parameters adaptively, have been proposed. One of these popular approaches is the heuristic optimization method. However, in terms of optimization, designing fuzzy logic control is a complex optimization problem that is discrete in terms of rule optimization and numerical in terms of optimization of membership degrees parameters. In this context, heuristic-based adaptive fuzzy control systems focus on either fuzzy rule optimization, weighting fuzzy rules, or parameter optimization. In this paper, unlike the others, an adaptive weighted fuzzy logic control (awFLC) method, which weights the inputs instead of the rules, is proposed. First, the membership degree of each input is calculated. Then, the resultant weight is determined by combining the weighted input membership degrees. For a crisp result, the average of the membership degrees of the resultant weight to the output membership functions is calculated. In awFLC, the interaction between membership functions is achieved by average membership degree, communication between inputs is achieved by the weighting of inputs, and mapping between inputs-outputs is achieved by the resultant weight value. Thus, the approach, which turns into a purely numerical optimization problem, provides convenience for heuristic search. In awFLC, optimal values for input weights and variable parameters are searched by the genetic algorithm. The performance of the method was tested on traction power control, and the results were compared with the ANFIS results. With awFLC, an 8.13% average error was obtained, while ANFIS produced solutions with an average error rate of 8.97%.

Ultrafast inverse design of quantum dot optical spectra via a joint TD-DFT learning scheme and deep reinforcement learning

Here, we report a case study on inverse design of quantum dot optical spectra using a deep reinforcement learning algorithm for the desired target optical property of semiconductor CdxSeyTex−y quantum dots. Machine learning models were trained to predict the optical absorption and emission spectra by using the training dataset by time dependent density functional theory simulation. We show that the trained deep deterministic policy gradient inverse design agent can infer the molecular structure with an accuracy of less than 1 Å at a fixed computational time of milliseconds and up to 100–1000 times faster than the conventional heuristic particle swam optimization method. Most of the effective inverse design problems based on the surrogate machine learning and reinforcement learning model have been focused on the field of nano-photonics. Few attempts have been made in the field of quantum optical system in a similar manner. For the first time, our results, to our knowledge, provide concrete evidence that for computationally challenging tasks, a well-trained deep reinforcement learning agent can replace the existing quantum simulation and heuristics optimization tool, enabling fast and scalable simulations of the optical property of nanometer sized semiconductor quantum dots.