metaheuristic algorithms

This paper introduces the Greedy Man Optimization Algorithm (GMOA), a novel bio-inspired metaheuristic approach for solving complex optimization problems. Inspired by competitive individuals resisting change, GMOA incorporates two unique mechanisms: MMO resistance, which prevents premature replacement of solutions, and periodic parasite removal, which promotes diversity and avoids stagnation. The algorithm is evaluated on standard benchmark functions, including Sphere, Rastrigin, Rosenbrock, and Griewank, and its performance is compared with established algorithms such as Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Differential Evolution (DE), and Ant Colony Optimization (ACO). Results demonstrate that GMOA outperforms these methods in terms of solution quality, convergence rate, and robustness. Statistical significance tests validate the reliability of the results. GMOA’s ability to balance exploration and exploitation makes it a promising tool for various real-world applications, including supply chain optimization and healthcare resource allocation.

Structural optimization plays a crucial role in engineering design, aiming to minimize weight and cost while satisfying performance constraints. This research presents a novel Self-Adaptive Enhanced Vibrating Particle System (SA-EVPS) algorithm that automatically adjusts algorithm parameters to improve optimization performance. The algorithm is applied to two challenging examples from the International Student Competition in Structural Optimization (ISCSO) benchmark suite: the 314-member truss structure (ISCSO_2018) and the 345-member truss structure (ISCSO_2021). Results demonstrate that SA-EVPS achieves significantly better solutions compared to previous studies using the Exponential Big Bang-Big Crunch (EBB-BC) algorithm. For ISCSO_2018, SA-EVPS achieved a minimum weight of 16543.57 kg compared to 17934.3 kg for the best EBB-BC variant—a 7.75% improvement. Similarly, for ISCSO_2021, SA-EVPS achieved 4292.71 kg versus 4399.0 kg for the best EBB-BC variant—a 2.42% improvement. The proposed algorithm also demonstrates superior convergence behavior and solution consistency, with coefficients of variation of 3.13% and 1.21% for the two benchmark problems, compared to 12.5% and 2.4% for the best EBB-BC variant. These results highlight the effectiveness of the SA-EVPS algorithm for solving complex structural optimization problems and demonstrate its potential for engineering applications.

In this study, the Improved Material Generation Algorithm (IMGA) is proposed to optimize the shape and size of structures. The original Material Generation Algorithm (MGA) introduced an optimization model inspired by the high-level and fundamental characteristics of material chemistry, particularly the configuration of compounds and chemical reactions for generating new materials. MGA uses a Gaussian normal distribution to produce new combinations. To enhance MGA for adapting truss structures, a new technique called Random Chaotic (RC) is proposed. RC increases the speed of convergence and helps escape local optima. To validate the proposed method, several truss structures, including a 37-bar truss bridge, a 52-bar dome, a 72-bar truss, a 120-bar dome, and a 200-bar planar structure, are optimized under natural frequency constraints. Optimizing the shape and size of structures under natural frequency constraints is a significant challenge due to its complexity. Choosing the frequency as a constraint prevents resonance in the structure, which can lead to large deformations and structural failure. Reducing the vibration amplitude of the structure decreases tension and deflection. Consequently, the weight of the structure can be minimized while keeping the frequencies within the permissible range. To demonstrate the superiority of IMGA, its results are compared with those of other state-of-the-art metaheuristic methods. The results show that IMGA significantly improves both exploitation and exploration.

Accurate estimation of bond strength between concrete and deformed reinforcing bars is essential for the stability of reinforced concrete structures, especially in critical regions subjected to heavy loads and environmental stresses. Despite intensive experimental studies revealing the complexity of factors influencing bond strength, existing predictive models, often reliant on artificial neural networks, have limitations in accuracy due to constrained datasets and inadequate representation of real-world stress fields. In response, this study pioneers a novel hybrid metaheuristic-optimized neural network model to swiftly and precisely predict bond strength under tensile load. Utilizing a comprehensive dataset comprising 558 valid experimental outcomes, seven metaheuristic algorithms are employed to optimize the ANN architecture. These metaheuristic algorithms include the Weighted Mean of Vectors, Grey Wolf Optimizer, Energy Valley Optimizer, Circle Search Algorithm, Artificial Ecosystem-Augmented Optimization, War Strategy Optimization, and Brown-Bear Optimization Algorithm. Results demonstrate that the developed hybrid models, particularly the artificial neural networks optimized by the Weighted Mean of Vectors algorithm, exhibit superior predictive performance. This model also demonstrated the lowest miscalibration value, followed by Circle Search Algorithm and Energy Valley Optimizer, indicating a high level of reliability. Moreover, comparison with common analytical and empirical formulations revealed significant performance improvements of the proposed model, achieving a 25% reduction in MSE during the testing phase. Additionally, the Shapley Additive explanations and Sobol sensitivity analysis framework was used to interpret the proposed predictive model, highlighting key predictors such as cross-sectional area, development length or splice, reinforcing bar diameter, and concrete compressive strength.

The aim of the present study is to predict audit failure using metaheuristic algorithms in companies listed on the Tehran Stock Exchange. To achieve this objective, 1,848 firm-year observations (154 companies over 12 years) were collected from the annual financial reports of companies listed on the Tehran Stock Exchange during the period from 2011 to 2022. In this study, four metaheuristic algorithms (including Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), and Bee Colony Optimization (BCO)) were utilized, as well as two methods for selecting the final research variables (the two-sample t-test and the forward stepwise selection method) to create the model. The results from the metaheuristic algorithms indicate that the overall accuracy of the GA, PSO, ACO, and BCO algorithms is 95.3%, 94.5%, 90.6%, and 92.8%, respectively, demonstrating the superiority of the Genetic Algorithm (GA) compared to other metaheuristic algorithms. Furthermore, the overall results from the variable selection methods indicate the efficiency of the stepwise method. Therefore, in companies listed on the Tehran Stock Exchange, the stepwise method and the Genetic Algorithm (GA) provide the most efficient model for predicting audit failure.

Sudden and severe stock price crashes pose a significant challenge to stock markets. The substantial losses incurred from such events underscore the need for more effective forecasting tools. This study aims to enhance the predictive power of models for stock price crashes in Tehran Stock Exchange and commenced with a comprehensive literature review to identify key financial factors influencing stock price volatility. Given the high dimensionality of the dataset and the extended time period, metaheuristic algorithms were employed for feature selection. 10 algorithms, namely Ant Colony Optimization, Hill Climbing, Las Vegas, Whale Optimization, Simulated Annealing, Genetic Algorithm, Tabu Search, Particle Swarm Optimization (PSO), Honey Bee (HBA) and Firefly were utilized to reduce dimensionality and enhance model performance. Subsequently, Support Vector Machines were implemented to develop predictive models. The models were trained and evaluated using historical data from Tehran Stock Exchange spanning from 2001 to 2020. The findings of this research demonstrate that combining metaheuristic algorithms for model reduction and optimization, along with advanced machine learning techniques, yields results that can significantly improve investment decision-making.

Nowadays, the project scheduling problem with limited resources has become one of the most important optimization issues. Given the enormous costs spent on projects as well as materials, resources, and forces, projects are expected to be successful and finish at the planned time. This schedule in functional mode will have many uncertainties due to the needs of the moment. Such a goal requires careful planning and advanced algorithms to solve these complex issues in a short and reasonable time. In this research, a new method is presented using the Intelligent Water Drops (IWD) algorithm for the resource-constrained project scheduling problem. For this reason and because of the importance of these projects, in this research, an optimization model has been developed for project scheduling in the state of uncertainty, which can solve many implementation obstacles. For this purpose, first, the problem is formulated as a mixed-inter linear programming (MILP) model. Next, the model is optimized using the IDW algorithm. To evaluate the performance of the proposed method, the standard data set was used in previous research and articles, and four datasets with different scales were selected from the PSLIB library. The results show that the proposed method is capable of obtaining the best precision in terms of the least critical deviation from the optimal solution. Moreover, the results of the proposed method were compared with metaheuristic algorithms, such as the particle congestion algorithm (PSO) , which was able to get the best solution among these algorithms.

A major design challenge for a grid-integrated photovoltaic power plant is to generate maximum power under varying loads, irradiance, and outdoor climatic conditions using competitive algorithm-based controllers. The objective of this study is to review experimentally validated advanced maximum power point tracking algorithms for enhancing power generation. A comprehensive analysis of 14 of the most advanced metaheuristics and 17 hybrid homogeneous and heterogeneous metaheuristic techniques is carried out, along with a comparison of algorithm complexity, maximum power point tracking capability, tracking frequency, accuracy, and maximum power extracted from PV systems. The results show that maximum power point tracking controllers mostly use conventional algorithms; however, metaheuristic algorithms and their hybrid variants are found to be superior to conventional techniques under varying environmental conditions. The Grey Wolf Optimization, in combination with Perturb & Observe, and Jaya-Differential Evolution, is found to be the most competitive technique. The study shows that standard testing and evaluation procedures can be further developed for comparing metaheuristic algorithms and their hybrid variants for developing advanced maximum power point tracking controllers. The identified algorithms are found to enhance power generation by grid-integrated commercial solar power plants. The results are of importance to the solar industry and researchers worldwide.

This study aimed to develop and optimize artificial stone mix designs incorporating microsilica using artificial neural networks (ANNs) and metaheuristic optimization algorithms. Initially, 10 base mix designs were prepared and tested based on previous experience and literature. The test results were used to train an ANN model. The trained ANN was then optimized using SA-EVPS and EVPS algorithms to maximize 28-day compressive strength, with aggregate gradation as the optimization variable. The optimized mixes were produced and tested experimentally, revealing some discrepancies with the ANN predictions. The ANN was retrained using the original and new experimental data, and the optimization process was repeated iteratively until an acceptable agreement was achieved between predicted and measured strengths. This approach demonstrates the potential of combining ANNs and metaheuristic algorithms to efficiently optimize artificial stone mix designs, reducing the need for extensive physical testing.

Great attention should be paid to planning and scheduling surgeries in the operating room which is the most sensitive ward in the health context in terms of cost and specific sensitivity due to its association with the life and death of individuals. In this case, reusable sterile equipment and devices are crucial issues because the hospital or nosocomial infections result from insufficient sterilization of these instruments. Therefore, sterilization of reusable medical devices is a necessity in the operating room to prevent possible infections. This study solves the integrated operating rooms and sterile section planning problem to minimize the total costs of sterilization, surgery postponement, and performance. This study also minimizes the completion time of surgery considering nondeterministic operating times and emergency-elective patients. In the real world, surgery time may be nondeterministic based on the conditions of the patient, surgeon, equipment, and instruments; hence, it is valuable to find a robust solution for planning under such circumstances. After presenting a bi-objective mathematical model for this problem, an improved epsilon constraint method was used to solve problems with small dimensions, and two metaheuristics NSGA-II and NRGA were developed for large dimensions regarding NP-hard problems. These two algorithms were analysed in terms of five indicators. The results indicated the superiority of the NSGA-II algorithm over NRGA to solve such problems.

The analysis and design of high-rise structures is one of the challenges faced by researchers and engineers due to their nonlinear behavior and large displacements. The moment frame system is one of the resistant lateral load-bearing systems that are used to solve this problem and control the displacements in these structures. However, this type of structural system increases the construction costs of the project. Therefore, it is necessary to develop a new method that can optimize the weight of these structures. In this work, the weight of these significant structures is optimized by using one of the latest metaheuristic algorithms called special relativity search. The special relativity search algorithm is mainly developed for the optimization of continuous unconstrained problems. Therefore, a penalty function is used to prevent violence of the constraints of the problem, which are tension, displacement, and drift. Also, using an innovative technique to transform the discrete problem into a continuous one, the optimal design is carried out. To prove the applicability of the new method, three different problems are optimized, including an eight-story one-span, a fifteen-story three-span bending frame, and a twenty-four-story three-span moment frame. The weight of the structure is the objective function, which should be minimized to the lowest possible value without violating the constraints of the problem. The calculation of stress and displacements of the structure is done based on the regulations of AISC-LRFD requirements. To validate, the results of the proposed algorithm are compared with other advanced metaheuristic methods.

A novel mixed integer non-linear mathematical model is presented in this paper for the two-echelon allocation-routing problem by applying the conditions of the route and transportation fleet under uncertainty. The cost of allocating drivers to non-homogeneous vehicles is calculated in this model based on the type of the vehicle, the lifecycle of the car, the experience of the driver, and different degrees of hardness that are defined for various routes. The cost of passing the route is defined based on an initial fixed cost and the degree of hardness of the route. Also, the reliability of the routes in each section is defined as an objective in the second echelon of the model aimed at enhancing the reliability rate. Two metaheuristic algorithms, NSGAII and MOPSO, are utilized to solve the model. Then, their performance rates in problems with different sizes are statistically evaluated and compared by different indices, following the adjustment of their parameters by Taguchi's method, through which results indicated the high efficiency of the model. A sensitivity analysis is ultimately performed on the results obtained from the solution, and some suggestions are made for the development of the model.

This paper presents the chaotic variants of the particle swarm optimization-statistical regeneration mechanism (PSO-SRM). The nine chaotic maps named Chebyshev, Circle, Iterative, Logistic, Piecewise, Sine, Singer, Sinusoidal, and Tent are used to increase the performance of the PSO-SRM. These maps are utilized instead of the random number, which defines the solution generation method. The robustness and performance of these methods are tested in the three steel frame design problems, including the 1-bay 10-story steel frame, 3-bay 15-story steel frame, and 3-bay 24-story steel frame. The optimization results reveal that the applied chaotic maps improve the performance of the PSO-SRM.

Structural topology optimization provides an insight into efficient designing as it seeks optimal distribution of material to minimize the total cost and weight of the structures. This paper presents an optimum design of steel moment frames and connections of structures subjected to serviceability and strength constraints in accordance with AISC-Load and Resistance Factor Design (LRFD). In connection topology optimizations, different beam and column sections and connections and also to optimize two steel moment frames a genetic algorithm was used and their performance was compared. Initially, two common steel moment frames were studied, only for the purpose of minimizing the weight of the structure and the members of structure are considered as design variables. Since the cost of a steel moment frame is not solely related to the weight of the structure, in order to obtain a realistic plan, in the second part of this study, for the other two frames the cost of the connections is also added to the variables. The results indicate that the steel frame optimization by applying real genetic algorithm could be optimal for structural designing. The findings highlighted the prominent performance and lower costs of the steel moment frames when different connections are used.

In this paper, three recently improved metaheuristic algorithms are utilized for the optimum design of the frame structures using the force method. These algorithms include enhanced colliding bodies optimization (ECBO), improved shuffled Jaya algorithm (IS-Jaya), and Vibrating particles system - statistical regeneration mechanism algorithm (VPS-SRM). The structures considered in this study have a lower degree of statical indeterminacy (DSI) than their degree of kinematical indeterminacy (DKI). Therefore, the force method is the most suitable analysis method for these structures. The robustness and performance of these methods are evaluated by the three design examples named 1-bay 10-story steel frame, 3-bay 15-story steel frame, and 3-bay 24-story steel frame.