Exploitation Capabilities Research Articles

Metaheuristic optimisation of Gaussian process regression model hyperparameters: Insights from FEREBUS

FEREBUS is a Gaussian process regression (GPR) engine embedded in the large machinery of FFLUX, a novel machine learnt force field developed from scratch through several well-documented proof-of-concept studies. This package relies on the exploration and exploitation capabilities of metaheuristic algorithms (MAs) to carry out the global optimisation of GPR model hyperparameters (θ). However, because MAs employ different search mechanisms to scrutinise the hyperparameter space, their performance on a specific optimisation task can vary a lot from one technique to another. Herein, we report a series of carefully designed experiments aimed at evaluating the ability of ten metaheuristic algorithms to locate the optimal set of θ values. Selected optimisation techniques belong to four popular families of MAs, namely particle swarm optimisation (4), grey wolf optimisation (2), bat (2) and firefly (2) algorithms. Our calculations suggest that grey wolf optimisers (GWOs) achieve the best results on average. Furthermore, the RMSE(θ) cost function is confirmed to be an excellent guide for the selection of atomic GPR models. This work also briefly introduces an enhanced grey wolf optimiser called GWO-RUHL (Random Update of the Hierarchy Ladder), which accounts for the (so far omitted) natural desire of non-leader wolves to occupy high-ranked leadership positions in the pack. We demonstrate that GWO-RUHL achieves better results than the standard GWO in terms of both convergence speed and quality of solutions.

Robust Q‐learning‐based multi‐objective sheep flock optimizer with a Cauchy operator for effective path planning in unmanned aerial vehicles

SummaryUnmanned aerial vehicle (UAV) path planning can be treated as a nondeterministic polynomial (NP) hard concern or an optimization problem. The conventional approaches are unable to effectively handle these issues due to discontinuity, non‐linearity, multi‐modality, and inseparability. On the other hand, meta‐heuristic algorithms are effective at tackling these issues because they are simple, adaptable, and derivation free. To enhance the performance in a variety of challenging circumstances, this paper proposes a novel Q‐learning‐based multi‐objective sheep flock optimizer with a Cauchy operator (Q‐MOSFO‐CA) to solve the constrained UAV path planning issues. The multi‐objective functions considered here are costs and constraints (threat, terrain, turning, climbing, and gliding constraints) to determine the feasible and optimal path. To avoid the probability of falling into the local optimum and to address the shortcoming of unbalanced convergence and also to maintain the exploitation and exploration capability, the Cauchy operator (CA) is integrated with the sheep flock optimization (SFO) algorithm. The Q‐learning model is introduced to balance both the global and local searches. Here, the exploration model performs the global search whereas the exploitation model performs the local search to attain an optimal solution. In the simulation scenario, the statistical analysis is conducted under two scenarios, and some essential measures such as the number of iterations at convergence (NIC), evaluation time (ET), energy consumption, and convergence analysis are determined. The proposed method obtains NIC of 1305 and 1436, ET of 12.8 and 15.2 s, and energy consumption of 20,600 and 21,465 J for both Scenarios 1 and 2, respectively.

A novel binary Kepler optimization algorithm for 0–1 knapsack problems: Methods and applications

The 0–1 Knapsack problem is a non-deterministic polynomial-time-hard combinatorial optimization problem that cannot be solved in reasonable time using traditional methods. Therefore, researchers have turned tometaheuristic algorithmsfor their ability to solve several combinatorial problems in a reasonable amount of time. This paper adapts theKepler optimization algorithm using eight V-shaped and S-shaped transfer functions to create abinary variantcalled BKOA for solving the 0–1 Knapsack problem. Several experiments were conducted to compare the efficacy of the binary Kepler optimization algorithm to severalcompeting optimizerswhen solving 20 well-known knapsack instances with dimensions ranging from 4 to 75. The experimental results demonstrate the superiority of this algorithm over other metaheuristic algorithms, except for thegenetic algorithm, which is marginally superior. To further improve the binary Kepler optimization algorithm, it is combined with an enhanced improvement strategy to create a new hybrid variant. This hybrid variant, termed HBKOA, has superior exploration and exploitation capabilities that make it better than genetic algorithm and other optimizers for all performance metrics considered. The enhanced improvement strategy is also integrated with several competing optimizers, and the experimental results show that HBKOA,hybrid binary manta ray foraging optimization, andhybrid binary equilibrium optimizer are competitive for small and medium-dimensional instances and superior for the higher dimensions.

IMPROVED MULTI-OBJECTIVE OPTIMIZATION IN BUSINESS PROCESS MANAGEMENT USING R-NSGA-II

Context. Business process management is a critical component in contemporary organizations for maintaining efficiency and achieving operational objectives. Optimization of these processes in terms of time and cost can lead to significant improvements in overall business performance. However, traditional optimization techniques often face challenges in handling multi-objective problems with a known time-cost trade-off, necessitating more effective solutions. The integration of a business process model and notation for a stochastic process simulation provides a robust foundation for analyzing these business processes and complies with stateof-the-art business process management. In prior studies, we applied several heuristic algorithms, including the evolutionary NSGAII, to find a Pareto-optimal set of solutions. We defined a solution as a pair of cost and time associated with a specific resource allocation. For one of the selected processes, the performance of NSGA-II was subpar compared to other techniques.
 Objective. The goal of this study is to improve upon the NSGA-II’s performance and, in turn, enhance the efficiency of multiobjective business process optimization. Specifically, we aim to incorporate reference points into NSGA-II. Our goal is to identify an optimized set of solutions that represent a trade-off between process execution time and the associated cost. We expect this set to have a higher spread and other quality metrics, compared to the prior outputs.
 Method. To accomplish our objective, we adopted a two-step approach. Firstly, we modified the original genetic algorithm by selecting and integrating the reference points that served to guide the search towards the Pareto-optimal front. This integration was designed to enhance the exploration and exploitation capabilities of the algorithm. Secondly, we employed the improved algorithm, namely R-NSGA-II, in the stochastic simulations of the business processes. The BPMN provided the input for these simulations, wherein we altered the resource allocation to observe the impact on process time and cost.
 Results. Our experimental results demonstrated that the R-NSGA-II significantly outperformed the original NSGA-II algorithm for the given process model, derived from the event log. The modified algorithm was able to identify a wider and more diverse Pareto-optimal front, thus providing a more comprehensive set of optimal solutions concerning cost and time.
 Conclusions. The study confirmed and underscored the potential of integrating the reference points into NSGA-II for optimizing business processes. The improved performance of R-NSGA-II, evident from the better Pareto-optimal front it identified, highlights its efficacy in multi-objective optimization problems, as well as the simplicity of the reference-based approaches in the scope of BPM. Our research poses the direction for the further exploration of the heuristics to improve the outcomes of the optimization techniques or their execution performance.

Q-learning based multi-objective immune algorithm for fuzzy flexible job shop scheduling problem considering dynamic disruptions

Confronted with complex industrial environments, dynamic disruptions like new job arrival and machine breakdown bring significant challenges to the robustness and stability of the manufacturing process, making the static production depart from the original scheduling scheme. To address this problem, a flexible job shop scheduling problem with fuzzy processing time, dynamic disruptions, and variable processing speeds is considered simultaneously. As well as three objectives of maximum completion time, total energy consumption, and average agreement index are demonstrated in this study. Then, a predictive-reactive dynamic/static rescheduling model is developed, where the off-line based mixed integer linear programming model and the on-line based rescheduling heuristics are proposed. Next, a multi-objective immune algorithm combined with a Q-learning algorithm (Q-MOIA) is developed. In the proposed algorithm, an active decoding heuristic based on the interval insertion mechanism is used to optimize the initial solutions. After that, the clonal selection-based immune algorithm and the Q-learning algorithm are adopted to improve the exploration and exploitation capabilities, respectively, where four objective-driven neighborhood structures are designed. Eventually, extensive computational experiments were conducted on 27 instances under static and dynamic scenarios to demonstrate the superiority and stability of the proposed predictive-reactive dynamic/static rescheduling model and the Q-MOIA. Comparative analysis with four state-of-the-art approaches revealed that proposed Q-MOIA outperformed in approximately 51.9, 66.7, and 83.3 % of the instances for the three multi-objective metrics.

A differential evolution algorithm combined with Lévy Flight for dimensional synthesis of four-bar linkage

The paper considers dimensional synthesis of a path generator four-bar linkage by applying the modified Differential Evolution (DE) algorithm. During the search of the space, the DE algorithm may skip the right solution and lead to premature convergence. If the step is reduced, the search space is reduced as well, so that convergence becomes slower. Search efficiency may be achieved by including Lévy Flight in the DE algorithm. Namely, when the DE algorithm identifies the space of best solution, then a Levý Flight step is used for local search in that region, which significantly increases the exploitation capability and allows obtaining a very efficient solution of the optimization problem. The algorithm was tested with three examples of dimensional synthesis of a path generator four-bar linkage. The proposed algorithm can be very efficiently applied to solving very complex optimization problems.

Genghis Khan shark optimizer: A novel nature-inspired algorithm for engineering optimization

This study tenders a new nature-inspired metaheuristic algorithm (MA) based on the behavior of the Genghis Khan shark (GKS), called GKS optimizer (GKSO), which is used for numerical optimization and engineering design. The inspiration for GKSO comes from the predation and survival behavior of GKS, and the entire optimization process is achieved by simulating four different activities of GKS, including hunting (exploration), movement (exploitation), foraging (switch from exploration to exploitation), and self-protection mechanism. These operators are mimicked using various mathematical models to efficiently perform optimization tasks of agents in different regions of the search space. In an effort to validate this method's viability and superiority, an in-depth analysis of the proposed GKSO is carried out from both qualitative and quantitative perspectives. Qualitative analysis verifies that GKSO has good exploration and exploitation (ENE) capability. Simultaneously, GKSO is quantitatively analyzed with eight existing fish optimization algorithms and the other nine well-known MAs on CEC2019 and CEC2022, respectively. Among them, a series of experimental scenarios are conducted to validate the applicability and robustness of GKSO by exploring its performance for CEC2022 at different dimensions and maximum fitness evaluation quantity. Statistical results indicate that GKSO has a strong advantage in the competition between two different types of algorithms. Furthermore, five different kinds of real-world constrained optimization problems (OPs) in CEC2020 benchmark constrained optimization functions, including 50 engineering case suites, are selected to evaluate GKSO's performance and the other seven optimizers, further validating GKSO's extensive usefulness and validity in solving practical complex problems.

A Nature Inspired Hybrid Partitional Clustering Method Based on Grey Wolf Optimization and JAYA Algorithm

This paper presents a hybrid meta-heuristic algorithm using Grey Wolf optimization (GWO) and JAYA algorithm for data clustering. The idea is use exploitative capability of JAYA algorithm in the explorative phase of GWO to form compact clusters. Here, instead of using one best and one worst solution for generating offspring, three best wolfs and three worst omega wolfs of the population are used. So, the best wolfs and worst omega wolfs assist in moving the new solutions towards the best solutions and simultaneously helps in staying away from the worst solutions. This enhances the chances of reaching the near optimal solutions. The superiority of the proposed method is compared with five promising algorithms, namely GWO, Sine-Cosine Algorithm (SCA), Particle Swarm Optimization (PSO), JAYA and K-means algorithms. The result obtained from the Duncan’s multiple range test and Nemenyi hypothesis based statistical test confirms the superiority and robustness of our proposed method.

Dynamic ensemble multi-strategy based bald eagle search optimization algorithm: A controller parameters tuning approach

To address the problems of bald eagle search algorithm (BES), easy fall into local optimums, limited diversity and slow convergence, a dynamic ensemble multi-strategy bald eagle search (DMBES) algorithm is proposed. Firstly, a nonlinear control factor is constructed to explore the potential range of population search, which helps to explore potential solutions. Next, a Levy flight strategy is introduced to better smooth the transition and explore the solution space through individual information interaction. Then, a dynamic selection strategy is proposed to help individuals obtain optimal solutions and better balance the algorithm's exploration and exploitation capabilities. Finally, in numerical experiments, classical benchmark functions and CEC 2022 test functions are selected to compare the performance of DMBES with other reliable and recent algorithms. In addition, box plots, Wilcoxon signed-rank test and Friedman rank test are used to confirm the significance of the results. In the controller parameter tuning experiments, the results show that the optimized controller of the DMBES algorithm outperforms the six comparison algorithms in terms of overshoot and rise time.

Enhanced Grey Wolf Optimization Algorithm for Mobile Robot Path Planning

In this study, an enhanced hybrid Grey Wolf Optimization algorithm (HI-GWO) is proposed to address the challenges encountered in traditional swarm intelligence algorithms for mobile robot path planning. These challenges include low convergence accuracy, slow iteration speed, and vulnerability to local optima. The HI-GWO algorithm introduces several key improvements to overcome these limitations and enhance performance. To enhance the population diversity and improve the initialization process, Gauss chaotic mapping is applied to generate the initial population. A novel nonlinear convergence factor is designed to strike a balance between global exploration and local exploitation capabilities. This factor enables the algorithm to effectively explore the solution space while exploiting the promising regions to refine the search. Furthermore, an adaptive position update strategy is developed by combining Levy flight and golden sine. This strategy enhances the algorithm’s solution accuracy, global search capability, and search speed. Levy flight allows longer jumps to explore distant regions, while golden sine guides the search towards the most promising areas. Extensive simulations on 16 standard benchmark functions demonstrate the effectiveness of the proposed HI-GWO algorithm. The results indicate that the HI-GWO algorithm outperforms other state-of-the-art intelligent algorithms in terms of optimization performance. Moreover, the performance of the HI-GWO algorithm is evaluated in a real-world path planning experiment, where a comparison with the traditional grey wolf algorithm and ant colony algorithm validates the superior efficiency of the improved algorithm. It exhibits excellent optimization ability, robust global search capability, high convergence accuracy, and enhanced robustness in diverse and complex scenarios. The proposed HI-GWO algorithm contributes to advancing the field of mobile robot path planning by providing a more effective and efficient optimization approach. Its improvements in convergence accuracy, iteration speed, and robustness make it a promising choice for various practical applications.

An Improved Differential Evolution for Parameter Identification of Photovoltaic Models

Photovoltaic (PV) systems are crucial for converting solar energy into electricity. Optimization, control, and simulation for PV systems are important for effectively harnessing solar energy. The exactitude of associated model parameters is an important influencing factor in the performance of PV systems. However, PV model parameter extraction is challenging due to parameter variability resulting from the change in different environmental conditions and equipment factors. Existing parameter identification approaches usually struggle to calculate precise solutions. For this reason, this paper presents an improved differential evolution algorithm, which integrates a collaboration mechanism of dual mutation strategies and an orientation guidance mechanism, called DODE. This collaboration mechanism adaptively assigns mutation strategies to different individuals at different stages to balance exploration and exploitation capabilities. Moreover, an orientation guidance mechanism is proposed to use the information of the movement direction of the population centroid to guide the evolution of elite individuals, preventing them from being trapped in local optima and guiding the population towards a local search. To assess the effectiveness of DODE, comparison experiments were conducted on six different PV models, i.e., the single, double, and triple diode models, and three other commercial PV modules, against ten other excellent meta-heuristic algorithms. For these models, the proposed DODE outperformed other algorithms, with the separate optimal root mean square error values of 9.86021877891317 × 10−4, 9.82484851784979 × 10−4, 9.82484851784993 × 10−4, 2.42507486809489 × 10−3, 1.72981370994064 × 10−3, and 1.66006031250846 × 10−2. Additionally, results obtained from statistical analysis confirm the remarkable competitive superiorities of DODE on convergence rate, stability, and reliability compared with other methods for PV model parameter identification.

Copula entropy-based golden jackal optimization algorithm for high-dimensional feature selection problems

Feature selection (FS) is a crucial process that aims to remove unnecessary features from datasets. It plays a role in data mining and machine learning (ML) by reducing the risk associated with high-dimensional datasets. FS is considered a challenging problem that is difficult to solve efficiently due to its combinatorial nature. As the size of the problem increases, the computation time also grows. Recently, researchers have focused on metaheuristic FS algorithms specifically designed for high-dimensional datasets. Therefore, this article proposes a powerful metaheuristic algorithm called Binary Enhanced Golden Jackal Optimization (BEGJO), which is an improved version of the recently published Golden Jackal Optimization (GJO) algorithm. The original GJO algorithm faces challenges when dealing with high-dimensional FS problems, as it tends to get trapped in local optima. To address this issue, various enhancement strategies are employed to improve the efficiency of GJO. The proposed BEGJO algorithm utilizes Copula Entropy (CE) to reduce the dimensionality of high-dimensional FS problems while maintaining high classification accuracy using the K-Nearest Neighbour (K-NN) classifier. Additionally, four enhancement strategies are incorporated to enhance the exploration and exploitation capabilities of the fundamental GJO algorithm. The BEGJO algorithm is transformed into its binary form using the sigmoid transfer function, aligning it with the nature of the FS problem. It is then tested on various high-dimensional benchmark datasets. The effectiveness of BEGJO is evaluated by comparing it with well-known algorithms in terms of classification accuracy, feature dimension, and processing time. BEGJO outperforms other algorithms in terms of classification accuracy and feature dimension and ranks up to fourth in terms of processing time. Furthermore, the advantageous use of CE is demonstrated by comparing the performance of the proposed algorithm with traditional FS algorithms. Statistical evaluations are conducted to further validate the effectiveness and superiority of the proposed algorithm. The results confirm that BEGJO is an effective solution for high-dimensional FS problems.

Probabilistic optimal planning of multiple photovoltaics and battery energy storage systems in distribution networks: A boosted equilibrium optimizer with time-variant load models

In recent years, there has been a rapid increase in distributed generation (DG) technologies incorporated into distribution networks (DNs) to meet the challenge of load growth. However, the stochastic nature of renewable energy, such as photovoltaic (PV), makes the amount of energy produced uncertain. This uncertainty, along with changes in generation and load demand, can increase energy losses and voltage instability. To address this issue, energy storage systems can be integrated to decrease the effects of the intermittency associated with renewable technologies. This paper proposes a new variant of an equilibrium optimizer (EO) based on reinforced learning, named RLEO, for optimal incorporation of multiple battery energy storage (BES) units integrated synchronously with solar PVs into distribution systems while minimizing energy loss. The RLEO algorithm employs reinforced learning mechanisms to prevent premature convergence of the EO and improve its exploration and exploitation capabilities. The performance of the RLEO algorithm is assessed using standard CEC 2017 benchmark functions and compared with the original EO and other popular algorithms using various statistical criteria. The RLEO algorithm is also applied to determine the optimal size and position of multiple PV units in IEEE 69-bus and 118-bus DNs with single and multi-objective optimization problems. Using the developed algorithm, the optimal arrangement of three non-dispatchable PVs results in a slight increase in the percentage reduction of energy loss across various load profiles: 53.0035 % for commercial, 19.6372 % for industrial, and 28.1783 % for residential. In contrast, by employing the optimal configuration of three PV + BES units, the reduction in energy loss percentage experiences a remarkable surge to 68.3466 % for commercial, 68.0917 % for industrial, and 68.1779 % for residential load scenarios using the proposed algorithm. This clearly indicates that the proposed RLEO algorithm significantly surpasses recent optimization methods documented in the literature when it comes to addressing the challenge of optimal allocation for multiple DGs. Furthermore, its applicability extends to more intricate optimization problems.

A multi-objective discrete differential evolution algorithm for energy-efficient distributed blocking flow shop scheduling problem

ABSTRACT The energy problem in green manufacturing has attracted enormous attention from researchers and practitioners in the manufacturing domain with the global energy crisis and the aggravation of environmental pollution. The distributed blocking flow shop scheduling problem (DBFSP) has considerable application scenarios in connection with its widespread application in the industry under the background of intelligent manufacturing. A multi-objective discrete differential evolution (MODE) algorithm is proposed to solve the energy-efficient distributed blocking flow shop scheduling problem (EEDBFSP) with the objectives of the makespan and total energy consumption (TEC) in this paper. The cooperative initialisation strategy is proposed to generate the initial population of the EEDBFSP. The mutation, crossover, and selection operators are redesigned to enable the MODE algorithm as applied to discrete space. A local search strategy based on the knowledge of five operators is introduced to enhance the exploitation capability of the MODE algorithm in the EEDBFSP. The non-critical path energy-efficient strategy is proposed to reduce energy consumption according to the specific constraints in the EEDBFSP. The effectiveness of each strategy in the MODE algorithm is verified and compared with the state-of-the-art algorithms. The numerical results demonstrate that the MODE algorithm is the efficient optimiser for solving the EEDBFSP.

A Gaussian random walk salp swarm algorithm for optimal dynamic charging of electric vehicles

Salp swarm algorithm (SSA) is one of the recently developed meta-heuristic optimization algorithms. Since SSA outperforms other swarm-based algorithms, it has recently been employed in various applications, including feature selection, neural network training and renewable energy systems. In this paper, an improved salp swarm algorithm based on a Gaussian random walk is proposed, which enhances the algorithm’s performance particularly for multidimensional constrained global optimization problems. The integration of a Gaussian random walk into the algorithm balances between its exploration and exploitation capabilities. Furthermore, the proposed algorithm introduces a new re-dispersion strategy in the case of stagnation at local optimum points, which considerably enhances exploration. The performance of the proposed algorithm is evaluated using a set of twenty-three benchmark test functions and is compared to the performance of prevalent metaheuristic algorithms. Statistical analysis is performed using Wilcoxon signed-rank test, and the results reveal considerable improvement over the competing algorithms. Then, 21 real-world optimization problems are used to further evaluate the efficacy of the proposed algorithm. The winners of the CEC2020 Competition on Real-World Single Objective Constrained Optimization, SASS, sCMAgES, EnMODE, and COLSHADE algorithms, are used as four comparable algorithms in the real-world optimization problems. The convergence curves and simulations provide very competitive performance compared to the comparative algorithms. The proposed algorithm is used to address one of the most challenging real-world constrained problems in power system applications, namely, determining the optimal charging schedule for electric vehicles at charging stations. The results reveal that the proposed algorithm outperforms other existing algorithms in terms of increasing the charging revenues and achieving maximum power grid stability.

A novel heuristic algorithm for solving engineering optimization and real-world problems: People identity attributes-based information-learning search optimization

With the scale and dimension of engineering optimization and real-world problems increasing, it will be difficult to find the optimum solutions. This paper proposes a novel people identity attributes-based heuristic technology, named the People Identity Attributes-based Information-learning Search Optimization (ISO), inspired by people’s psychological assessment and learning behaviors for information resources based on the social status-based identity attribute and the self-survival demands-based identity attribute. For the former, a four-level information delivery mechanism is constructed through leader, manager, executor, and freelancer, emphasizing global optimization. For the latter, the staged transforming model based on random accumulation and directed induction behaviors are constructed according to self-survival demands, emphasizing global optimization and exploration. The dual identity attributes-based search strategy emphasizes psychological assessment and selection for the information loss, including the learning behavior for two kinds of information resources with different identity attributes, and the recovering behavior for the information loss by constructing the information resilience equation, which emphasizes local optimization and exploitation. This paper qualitatively analyzes the swarm behavior, search history, and the exploration and exploitation capabilities of ISO. The optimization performances are quantitatively analyzed for ISO and 9 competitive algorithms on 39 benchmark tests, including the convergence, solution accuracy, robustness, sensitivity, significance, statistical investigation-based Wilcoxon test and Friedman test. The scalability of ISO is investigated on CEC2017 (30Dim, 50Dim, 100Dim) and the latest CEC2022 (10Dim, 20Dim) suites. The results reveal that compared to other competitive algorithms, ISO possesses best computing performance with ranking the first in all competitors. In addition, the proposed ISO and 12 competitors consider 10 constraint engineering optimization problems and the real application of path planning with multiple obstacles, suggesting that ISO possesses significant optimization performance.

HSFA: A novel firefly algorithm based on a hierarchical strategy

The firefly algorithm (FA) is a well-known metaheuristic algorithm that has demonstrated effectiveness and superiority in a variety of optimization tasks. To alleviate the oscillation caused by fireflies during the search process and improve the solution quality of the FA, this paper proposes an improved FA based on the hierarchy strategy (HSFA). In the HSFA, the firefly population was separated into elite and non-elite groups based on fitness values, and distinct attractive models were used for each group to improve the performance of the HSFA from the perspective of exploration and exploitation capabilities. This enhances the convergence speed and solution quality of the algorithm. The HSFA was run on two sets of standard benchmark functions, and its performance was evaluated using two practical engineering design problems. The experimental results demonstrate that the HSFA outperforms state-of-the-art optimization algorithms.

An embedded feature selection approach for depression classification using short text sequences

Depression has become a serious mental health issue worldwide, particularly due to the rise of the Global Pandemic. Identifying depression of an individual from short texts shared in social media is a challenging task. The present work aims to select the optimal feature subset for classifying short texts for depression detection. By performing feature selection, it is possible to eliminate redundant and noisy features in high-dimensional datasets with small sample sizes. This can prevent the ”curse of dimensionality” and enhance the effectiveness of classification algorithms. However, current feature selection methods often focus on optimizing classification or clustering performance, while neglecting the stability of the selected features. This can lead to unstable results and make it challenging to identify meaningful and interpretable features. This paper introduces a novel embedded feature selection approach named Statistical Relevance Class Frequency based on Whale Optimization Algorithm (SRCF-WOA) for selecting feature subsets from short texts in social media. The proposed methodology extracts both the unigram features and composite features to capture the semantic and structural information. χ2.rcf (Chi-squared relevance class frequency) filter approach is applied to rank the extracted features to signify the importance of the features. WOA is adapted to retrieve the optimal subset of features with low-dimensional space using its high exploration and high exploitation capability. In the evaluation process, four benchmark short text datasets and two classifiers are used. The comparison shows that the proposed embedded feature selection method outperforms other algorithms in terms of accuracy and Fβ scores(β=0.5,1, and 2). The sensitivity analysis is carried out to check the robustness and stability of the proposed method. The findings indicate that the SRCF-WOA surpasses other methods on the majority of datasets, achieving the maximum classification accuracy while utilizing the minimal features. The statistical importance of these findings is further supported by the Analysis of Variance (ANOVA) F-test. Moreover, the proposed method strikes the optimal balance between classification accuracy and feature stability.

A robust multiple Unmanned Aerial Vehicles 3D path planning strategy via improved particle swarm optimization

Path planning algorithms for Unmanned Aerial Vehicles (UAVs) are essential in various domains, such as search and rescue operations, agriculture, and delivery services. Nonetheless, finding optimal flight paths in complex environments with obstacles remains challenging. This paper proposes a novel path planning algorithm that uses a Nash equilibrium approach based on game theory to strike a balance between the exploitation and exploration capabilities of UAVs. The algorithm is designed in a simulated flight environment with obstacles, utilizing the self-awareness and group awareness coefficients from particle swarm optimization(PSO) as participants in the game theory model. The proposed path planning algorithm significantly improves the navigation of multiple UAVs in complex scenarios with obstacles by achieving a balance between their exploitation and exploration capabilities. Simulation experiments demonstrate that the proposed method surpasses traditional approaches, exhibiting an average improvement of 32.57%, 32.17%, and 29.33% in the algorithm convergence time, average flight distance, and average flight time, respectively. The significance of this research lies in its contribution to advancing UAV path planning algorithms. By integrating game theory and PSO, the proposed method optimizes the exploration and exploitation capabilities of UAVs, leading to improved navigation and resource utilization.

A novel collaborative iterative greedy algorithm for hybrid flowshop scheduling problem with batch processing machines and variable sublots

ABSTRACT Lot streaming technology enables continuous overlapping operations, which is of great significance in shortening production cycles, reducing unnecessary waiting time, and increasing production capacity. However, the capacity constraint of batch processing machines may lead to inevitable variations in sublots. Therefore, the key focus of our research is to control variations of sublot for maximising benefits. In view of this, we investigate a hybrid flowshop scheduling problem (HFSP) with batch processing machines and variable sublots (HFSP-BVS) integrating sequence-dependent setup times and transportation times. To address HFSP-BVS, a MILP model is first established, and a novel collaborative iterative greedy (NCIG) algorithm is proposed to optimise the cumulative payoffs associated with delivery dates. In NCIG, a collaborative initialisation method by extracting good information from an archive is proposed, and a specific destruction-reconfiguration strategy is designed to control the variations of sublots in the batch processing stage. Furthermore, a dynamic acceptance criterion is designed to balance the algorithm's exploitation and exploration capabilities. Lastly, we conduct comparisons between the NCIG algorithm and five other metaheuristic algorithms on 100 test instances. The results show that NCIG outperforms them by 1.89% and 61.42% on average in terms of the total penalty and RPI values, respectively.