Model Free Approach to Adaptive Parameters Optimisation using Adaptive Decaying Merit Score (ADMS): A Preliminary Study

Q-Learning-based parameter control in differential evolution for structural optimization

Differential evolution with orthogonal array‐based initialization and a novel selection strategy

Opposition-based Magnetic Optimization Algorithm with parameter adaptation strategy

This study addresses the challenges of real-time channel state estimation and adaptive parameter adjustment in dynamic LoRa networks, where the existing methods often fail to adapt efficiently to highly variable channel conditions. This study presents an innovative approach for real-time channel state estimation and adaptive parameter adjustment in long-range (LoRa) networks in dynamic Internet of Things (IoT) environments. When these types of networks are used in dynamic IoT environments, they are known to face challenges in the two above-mentioned areas. In our approach, a hybrid feature extraction method that integrates statistical analysis with domain-specific knowledge is utilized for real-time data labeling, focusing on the signal-to-noise (SNR) and received signal strength indicator (RSSI) metrics. This approach employs an adaptive sliding window technique for efficient processing of recent data. Subsequently, a multi-task long short-term memory (LSTM) neural network is introduced for the simultaneous prediction of multiple channel states. This multi-task model employs an online incremental learning approach to enhance the real-time performance and responsiveness of the model within dynamic environments. It also incorporates a confidence measure for estimated states to increase the prediction reliability. Finally, based on the confidence measure predictions and channel state estimation, the system dynamically adjusts the LoRa parameters, including the spreading factor, coding rate, transmission power, and bandwidth. Our results demonstrate that the confidence-based adaptive strategy coupled with adaptive sliding window processing and incremental learning effectively balances performance optimization with stability in challenging IoT scenarios. This study contributes a robust, data-driven approach for real-time channel state estimation and adaptive parameter control, addressing the unique challenges of IoT networks in dynamic environments. Our approach achieved a packet delivery ratio of 100%, reduced energy consumption to 0.07987 Joules per packet, and demonstrated a prediction accuracy between 97.70% and 97.9% for estimating the different channel states. This innovative framework provides significant improvements in channel state estimation, communication reliability, adaptive parameter control, and computational efficiency, thereby ensuring robust performance in IoT environments at the same time.

In this paper we proposed the circle trajectory assembly algorithm to control the multi-UAVs circular assembly formation. CTFAP solution provides rapid formation of UAVs on a circular orbit and solve the problem of large scattering distance. Proposed Distributed model prediction control framework improves the optimization ability and reduces the computation consumption with the better convergence ability of the UAV formation. Firstly, a circular trajectory following algorithm with an adaptive parameter is proposed to complete the rapid formation of UAVs on a circular orbit and solve the problem of large scattering distance during formation forming. Then, in the stage of formation reconfiguration, with distributed model prediction control framework (DMPC), the proposed method gets the prediction information of DMPC to optimize the population of classical differential evolution (DE) algorithm and improve the iterative optimization ability of DE algorithm. Experiments show that the proposed differential evolution algorithm greatly improves the efficiency of solving the formation reconfiguration problem under the DMPC framework and overcomes the disadvantages of random population of classical DE. For the proposed rapid forming method, assembly range is reduced by 41% compared with direct linearly formation assembly, and the formation forming time is reduced by approximately 21%. Compared with other optimization algorithms such as Particle Swarm Optimization (PSO), Genetic Algorithm (GA) and DE, the proposed differential evolution algorithm reduces instruction response time of single-drones by 16%–30%.

In order to solve problems with the original crayfish optimization algorithm (COA), such as reduced diversity, local optimization, and insufficient convergence accuracy, a multi-strategy optimization algorithm for crayfish based on differential evolution, named the ICOA, is proposed. First, the elite chaotic difference strategy is used for population initialization to generate a more uniform crayfish population and increase the quality and diversity of the population. Secondly, the differential evolution strategy and the dimensional variation strategy are introduced to improve the quality of the crayfish population before its iteration and to improve the accuracy of the optimal solution and the local search ability for crayfish at the same time. To enhance the updating approach to crayfish exploration, the Levy flight strategy is adopted. This strategy aims to improve the algorithm's search range and local search capability, prevent premature convergence, and enhance population stability. Finally, the adaptive parameter strategy is introduced to improve the development stage of crayfish, so as to better balance the global search and local mining ability of the algorithm, and to further enhance the optimization ability of the algorithm, and the ability to jump out of the local optimal. In addition, a comparison with the original COA and two sets of optimization algorithms on the CEC2019, CEC2020, and CEC2022 test sets was verified by Wilcoxon rank sum test. The results show that the proposed ICOA has strong competition. At the same time, the performance of ICOA is tested against different high-performance algorithms on 6 engineering optimization examples, 30 high-low-dimension constraint problems and 2 large-scale NP problems. Numerical experiments results show that ICOA has superior performance on a range of engineering problems and exhibits excellent performance in solving complex optimization problems.

The performance of differential evolution (DE) is influenced by various factors such as algorithm parameters and mutation strategies. JADE (adaptive DE with optional external archive) is recognized as one of the most successful studies on parameter control in DE. In this study, in order to improve the search efficiency of JADE, we propose to integrate two methods we have proposed. One is the method to estimate whether the population of candidate solutions is converging or moving for particle swarm optimization. The normalized distance between the population center and the best solution is used for the estimation. When the population is converging, the search mode is adjusted to enhance the convergence. When the population is moving, the search mode is adjusted to enhance the movement. The other is the method to control DE parameters for extreme individuals in order to improve the search efficiency. The methods are modified to use the former method in JADE and to reduce the interference between the two methods. The effectiveness of JADE incorporating the proposed methods is shown by solving thirteen benchmark problems.

The research presented in this article focuses on the development of a multi-objective optimization algorithm based on the differential evolution (DE) concept combined with Mamdani-type fuzzy logic controllers (FLCs) and $$K$$ K -medoids clustering. The FLCs are used for adaptive control of the DE parameters; $$K$$ K -medoids clustering enables the algorithm to perform a more guided search by evolving neighboring vectors, i.e., vectors that belong to the same cluster. A modified version of the $$DE/best/1/bin$$ D E / b e s t / 1 / b i n algorithm is adopted as the core search component of the multi-objective optimizer. The FLCs utilize Pareto dominance and cluster-related information as input in order to adapt the algorithmic parameters dynamically. The proposed optimization algorithm is tested using a number of problems from the multi-objective optimization literature in order to investigate the effect of clustering and parameter adaptation on the algorithmic performance under various conditions, e.g., problems of high dimensionality, problems with non-convex Pareto fronts, and problems with discontinuous Pareto fronts. A detailed performance comparison between the proposed algorithm with state-of-the-art multi-objective optimizers is also presented.

This research deals with the hybridization of several approaches for evolutionary algorithms, which are the adaptive control parameter adjusting strategy and multi-chaotic dynamics driving the selection of indices in Differential Evolution (DE). The novelty of the paper is given by the experiments with the multi-chaos-driven adaptive DE concept inside adaptive parameter adjusting DE strategies. These experiments are representing the investigations on the mutual influences of several different randomizations types together with adaptive DE strategies. The multi-chaotic concept is representing the adaptive switching between two different chaotic systems based on the progress of individuals within population. This paper is aimed at the embedding of discrete dissipative chaotic systems in the form of multi-chaotic pseudo random number generators for the jDE, which is the state of the art representative of simple adaptive control parameter adjusting strategy for DE. Repeated simulations for two different combinations of driving chaotic systems were performed on the IEEE CEC 13 benchmark set. Finally, the obtained results are compared with the canonical not-chaotic jDE.

An adaptive differential evolution with opposition-learning based diversity enhancement

Differential evolution (DE) is one of the most successful evolutionary algorithms (EAs) for global numerical optimization. Like other EAs, maintaining population diversity is important for DE to escape from local optima and locate a near-global optimum. Using a multi-population algorithm is a representative method to avoid early loss of population diversity. In this paper, we propose a multi-population DE algorithm (MPDE) which manipulates multiple sub-populations. Different sub-populations in MPDE exchange information via a novel mutation operation instead of migration used in most multi-population EAs. The mutation operation is helpful to balance the fast convergence and population diversity of the proposed algorithm. Moreover, the performance of MPDE is further improved by an adaptive parameter control scheme designed based on the multi-population approach. Each sub-population in MPDE evolves with its own set of control parameters, and a learning strategy is used to adaptively adjust the parameter values. A set of benchmark functions is used to test the proposed MPDE algorithm. The experimental results show that MPDE performs better than, or at least comparably, to the classical single population DE with fixed parameter values and three existing state-of-the-art DE variants.

The parameter identification and adaptive control schemes applied in a previous study to a nonlinear point reactor are extended to the case of a loosely-coupled-core reactor with internal feedbacks, constituting a nonlinear overall system. Both schemes are shown to be stable, with the system newly represented on the pattern of the Model Reference Adaptive System (MRAS) with use made of the Lyapunov's method. For either parameter identification or adaptive control of a loosely-coupled-core reactor, there exists no canonical form of multiple input-multiple output system which can be directly applied for deriving the MRAS with the matrix version of the Kalman-Yakubovich lemma as it was in the case of the point reactor. This difficulty is circumvented by the practical assumption that the neutron density can be directly measured on each core as reactivity change is applied as input into the coupled core as a whole. For parameter identification, the model parameters are adaptively adjusted to those of each core, while for the adaptive control, plant parameters of each core can be adaptively compensated, again through control inputs, to asymptotically reduce the output error between the model and the plant. The point reactor is shown to correspond to a special case. KEYWORDS: parameter identificationadaptive controlloosely-coupled-core reactorinternal feedbacks, nonlinearmodel reference adaptive systemLyapunov functionstability

Differential evolution (DE) is a widely used optimization algorithm, which can achieve high accuracy with a simple mechanism, but sometimes have only limited performances due to its simplicity. In order to mitigate the inappropriate effect of poor initial search points, it is known that adding a random search to DE contributes to obtain better results than normal DE. However, it is inefficient to perform many random searches when the search process is almost converged. In this study, we propose a novel method of DE with Adaptive Randomness (DEAR), which is a hybrid of two promising algorithms of DIEtoDE and SaDE, and can adaptively change the frequency of random search maintaining efficiency. Numerical experiments demonstrated that the proposed method can identify better solutions than other comparative methods.

The design of a new hybrid controller for fractional-order uncertain chaotic systems with unknown time-varying delays

In modern aviation power systems, the dynamic stability of power generation systems has become a major challenge. To tackle issues such as slow response, poor voltage stability, and voltage oscillations in aircraft synchronous generators under sudden load changes, a Proportional-Integral-Derivative (PID) voltage regulator modified using the Differential Evolution (DE) algorithm has been developed. Additionally, the DE algorithm’s convergence speed and robustness are enhanced through the incorporation of directional search strategies and adaptive control parameters. Simulation results demonstrate that the optimized regulator delivers rapid response and minimal overshoot, thereby improving the accuracy and dynamic performance of the voltage regulator. The modified DE algorithm notably accelerates convergence compared to the traditional DE algorithm.