Showing papers on "Metaheuristic published in 2020"

Equilibrium optimizer: A novel optimization algorithm

Abstract: This paper presents a novel, optimization algorithm called Equilibrium Optimizer (EO), inspired by control volume mass balance models used to estimate both dynamic and equilibrium states. In EO, each particle (solution) with its concentration (position) acts as a search agent. The search agents randomly update their concentration with respect to best-so-far solutions, namely equilibrium candidates, to finally reach to the equilibrium state (optimal result). A well-defined “generation rate” term is proved to invigorate EO’s ability in exploration, exploitation, and local minima avoidance. The proposed algorithm is benchmarked with 58 unimodal, multimodal, and composition functions and three engineering application problems. Results of EO are compared to three categories of existing optimization methods, including: (i) the most well-known meta-heuristics, including Genetic Algorithm (GA), Particle Swarm Optimization (PSO); (ii) recently developed algorithms, including Grey Wolf Optimizer (GWO), Gravitational Search Algorithm (GSA), and Salp Swarm Algorithm (SSA); and (iii) high performance optimizers, including CMA-ES, SHADE, and LSHADE-SPACMA. Using average rank of Friedman test, for all 58 mathematical functions EO is able to outperform PSO, GWO, GA, GSA, SSA, and CMA-ES by 60%, 69%, 94%, 96%, 77%, and 64%, respectively, while it is outperformed by SHADE and LSHADE-SPACMA by 24% and 27%, respectively. The Bonferroni–Dunnand Holm’s tests for all functions showed that EO is significantly a better algorithm than PSO, GWO, GA, GSA, SSA and CMA-ES while its performance is statistically similar to SHADE and LSHADE-SPACMA. The source code of EO is publicly availabe at https://github.com/afshinfaramarzi/Equilibrium-Optimizer , http://built-envi.com/portfolio/equilibrium-optimizer/ and http://www.alimirjalili.com/SourceCodes/EOcode.zip .

Chimp optimization algorithm

Abstract: This paper proposes a novel metaheuristic algorithm called Chimp Optimization Algorithm (ChOA) inspired by the individual intelligence and sexual motivation of chimps in their group hunting, which is different from the other social predators. ChOA is designed to further alleviate the two problems of slow convergence speed and trapping in local optima in solving high-dimensional problems. In this paper, a mathematical model of diverse intelligence and sexual motivation of chimps is proposed. In this regard, four types of chimps entitled attacker, barrier, chaser, and driver are employed for simulating the diverse intelligence. Moreover, four main steps of hunting, i.e. driving, chasing, blocking, and attacking, are implemented. The proposed ChOA algorithm is evaluated in 3 main phases. First, a set of 30 mathematical benchmark functions is utilized to investigate various characteristics of ChOA. Secondly, ChOA was tested by 13 high-dimensional test problems. Finally, 10 real-world optimization problems were used to evaluate the performance of ChOA. The results are compared to several newly proposed meta-heuristic algorithms in terms of convergence speed, the probability of getting stuck in local minimums, and exploration, exploitation. Also, statistical tests were employed to investigate the significance of the results. The results indicate that the ChOA outperforms the other benchmark optimization algorithms.

A test-suite of non-convex constrained optimization problems from the real-world and some baseline results

Abstract: Real-world optimization problems have been comparatively difficult to solve due to the complex nature of the objective function with a substantial number of constraints. To deal with such problems, several metaheuristics as well as constraint handling approaches have been suggested. To validate the effectiveness and strength, performance of a newly designed approach should be benchmarked by using some complex real-world problems, instead of only the toy problems with synthetic objective functions, mostly arising from the area of numerical analysis. A list of standard real-life problems appears to be the need of the time for benchmarking new algorithms in an efficient and unbiased manner. In this study, a set of 57 real-world Constrained Optimization Problems (COPs) are described and presented as a benchmark suite to validate the COPs. These problems are shown to capture a wide range of difficulties and challenges that arise from the real life optimization scenarios. Three state-of-the-art constrained optimization methods are exhaustively tested on these problems to analyze their hardness. The experimental outcomes reveal that the selected problems are indeed challenging to these algorithms, which have been shown to solve many synthetic benchmark problems easily.

Population size in Particle Swarm Optimization

Abstract: Particle Swarm Optimization (PSO) is among the most universally applied population-based metaheuristic optimization algorithms. PSO has been successfully used in various scientific fields, ranging from humanities, engineering, chemistry, medicine, to advanced physics. Since its introduction in 1995, the method has been widely investigated, which led to the development of hundreds of PSO versions and numerous theoretical and empirical findings on their convergence and parameterization. However, so far there is no detailed study on the proper choice of PSO swarm size, although it is widely known that population size crucially affects the performance of metaheuristics. In most applications, authors follow the initial suggestion from 1995 and restrict the population size to 20–50 particles. In this study, we relate the performance of eight PSO variants to swarm sizes that range from 3 up to 1000 particles. Tests are performed on sixty 10- to 100-dimensional scalable benchmarks and twenty-two 1- to 216-dimensional real-world problems. Although results do differ for the specific PSO variants, for the majority of considered PSO algorithms the best performance is obtained with swarms composed of 70–500 particles, indicating that the classical choice is often too small. Larger swarms frequently improve efficiency of the method for more difficult problems and practical applications. For unimodal problems slightly lower swarm sizes are recommended for the majority of PSO variants, but some would still perform best with hundreds of particles.

A better balance in metaheuristic algorithms: Does it exist?

Abstract: The constant development of new metaheuristic algorithms has led to a saturation in the field of stochastic search. There are now hundreds of different algorithms that can be used to solve any problem. To produce a good performance, every metaheuristic method needs to address a satisfactory equilibrium between exploration and exploitation of the search space. Although exploration and exploitation represent two fundamental concepts in metaheuristics, the main questions about their combination and balance have not been yet completely understood. Most of the existent analyzes conducted on metaheuristic techniques consider only the comparison of their final results which cannot evaluate the nature of a good or bad balance. This paper presents an experimental analysis that quantitatively evaluates the balance between exploration and exploitation of several of the most important and better-known metaheuristic algorithms. In the study, a dimension-wise diversity measurement is used to assess the balance of each scheme considering a representative set of 42 benchmark problems that involve multimodal, unimodal, composite and shifted functions. As a result, the analysis provides several observations that allow understanding how this balance affects the results in each type of functions, and which balance is producing better solutions.

Boosting salp swarm algorithm by sine cosine algorithm and disrupt operator for feature selection

Abstract: Features Selection (FS) plays an important role in enhancing the performance of machine learning techniques in terms of accuracy and response time. As FS is known to be an NP-hard problem, the aim of this paper is to introduce basically a new variant of Salp Swarm Optimizer (SSA) for FS (called ISSAFD (Improved Followers of Salp swarm Algorithm using Sine Cosine algorithm and Disrupt Operator), that updates the position of followers (F) in SSA using sinusoidal mathematical functions that were inspired from the Sine Cosine Algorithm (SCA). This enhancement helps to improve the exploration phase and to avoid stagnation in a local area. Moreover, the Disruption Operator (Dop) is applied for all solutions, in order to enhance the population diversity and to maintain the balance between exploration and exploitation processes. Two other variants of SSA are developed based on SCA called ISSALD (Improved Leaders of Salp swarm Algorithm using Sine Cosine algorithm and Disrupt Operator) and ISSAF (Improved Followers of Salp swarm Algorithm using Sine Cosine algorithm). The updating process in consists to update the leaders (L) position by SCA and applying (Dop), whereas in ISSAF, the Dop is omitted and the position of followers is updated by SCA. Experimental results are evaluated on twenty datasets where four of them represent high dimensionality with a small number of instances. The obtained results show a good performance of ISSAFD in terms of accuracy, sensitivity, specificity, and the number of selected features in comparison with other metaheuristics (MH).

An enhanced associative learning-based exploratory whale optimizer for global optimization

Abstract: Whale optimization algorithm (WOA) is a recent nature-inspired metaheuristic that mimics the cooperative life of humpback whales and their spiral-shaped hunting mechanism. In this research, it is first argued that the exploitation tendency of WOA is limited and can be considered as one of the main drawbacks of this algorithm. In order to mitigate the problems of immature convergence and stagnation problems, the exploitative and exploratory capabilities of modified WOA in conjunction with a learning mechanism are improved. In this regard, the proposed WOA with associative learning approaches is combined with a recent variant of hill climbing local search to further enhance the exploitation process. The improved algorithm is then employed to tackle a wide range of numerical optimization problems. The results are compared with different well-known and novel techniques on multi-dimensional classic problems and new CEC 2017 test suite. The extensive experiments and statistical tests show the superiority of the proposed BMWOA compared to WOA and several well-established algorithms.

Group teaching optimization algorithm: A novel metaheuristic method for solving global optimization problems

Abstract: In last 30 years, many metaheuristic algorithms have been developed to solve optimization problems. However, most existing metaheuristic algorithms have extra control parameters except the essential population size and stopping criterion. Considering different characteristics of different optimization problems, how to adjust these extra control parameters is a great challenge for these algorithms in solving different optimization problems. In order to address this challenge, a new metaheuristic algorithm called group teaching optimization algorithm (GTOA) is presented in this paper. The proposed GTOA is inspired by group teaching mechanism. To adapt group teaching to be suitable for using as an optimization technique, without loss of generality, four simple rules are first defined. Then a group teaching model is built under the guide of the four rules, which consists of teacher allocation phase, ability grouping phase, teacher phase and student phase. Note that GTOA needs only the essential population size and stopping criterion without extra control parameters, which has great potential to be used widely. GTOA is first examined over 28 well-known unconstrained benchmark problems and the optimization results are compared with nine state-of-the-art algorithms. Experimental results show the superior performance of the proposed GTOA for these problems in terms of solution quality, convergence speed and stability. Furthermore, GTOA is used to solve four constrained engineering design optimization problems in the real world. Simulation results demonstrate the proposed GTOA can find better solutions with faster speed compared with the reported optimizers.

A WOA-Based Optimization Approach for Task Scheduling in Cloud Computing Systems

Abstract: Task scheduling in cloud computing can directly affect the resource usage and operational cost of a system. To improve the efficiency of task executions in a cloud, various metaheuristic algorithms, as well as their variations, have been proposed to optimize the scheduling. In this article, for the first time, we apply the latest metaheuristics whale optimization algorithm (WOA) for cloud task scheduling with a multiobjective optimization model, aiming at improving the performance of a cloud system with given computing resources. On that basis, we propose an advanced approach called I mproved W OA for C loud task scheduling (IWC) to further improve the optimal solution search capability of the WOA-based method. We present the detailed implementation of IWC and our simulation-based experiments show that the proposed IWC has better convergence speed and accuracy in searching for the optimal task scheduling plans, compared to the current metaheuristic algorithms. Moreover, it can also achieve better performance on system resource utilization, in the presence of both small and large-scale tasks.

Selective Opposition based Grey Wolf Optimization

Abstract: The use of metaheuristics is widespread for optimization in both scientific and industrial problems due to several reasons, including flexibility, simplicity, and robustness. Grey Wolf Optimizer (GWO) is one of the most recent and popular algorithms in this area. In this work, opposition-based learning (OBL) is combined with GWO to enhance its exploratory behavior while maintaining a fast convergence rate. Spearman's correlation coefficient is used to determine the omega (ω) wolves (wolves with the lowest social status in the pack) on which to perform opposition learning. Instead of opposing all the dimensions in the wolf, a few dimensions of the wolf are selected on which opposition is applied. This assists with avoiding unnecessary exploration and achieving a fast convergence without deteriorating the probability of finding optimum solutions. The proposed algorithm is tested on 23 optimization functions. An extensive comparative study demonstrates the superiority of the proposed method. The source code for this algorithm is available at "https://github.com/dhargupta-souvik/sogwo"

Elephant Herding Optimization: Variants, Hybrids, and Applications

Abstract: Elephant herding optimization (EHO) is a nature-inspired metaheuristic optimization algorithm based on the herding behavior of elephants. EHO uses a clan operator to update the distance of the elephants in each clan with respect to the position of a matriarch elephant. The superiority of the EHO method to several state-of-the-art metaheuristic algorithms has been demonstrated for many benchmark problems and in various application areas. A comprehensive review for the EHO-based algorithms and their applications are presented in this paper. Various aspects of the EHO variants for continuous optimization, combinatorial optimization, constrained optimization, and multi-objective optimization are reviewed. Future directions for research in the area of EHO are further discussed.

Multi-Sequence LSTM-RNN Deep Learning and Metaheuristics for Electric Load Forecasting

Abstract: Short term electric load forecasting plays a crucial role for utility companies, as it allows for the efficient operation and management of power grid networks, optimal balancing between production and demand, as well as reduced production costs. As the volume and variety of energy data provided by building automation systems, smart meters, and other sources are continuously increasing, long short-term memory (LSTM) deep learning models have become an attractive approach for energy load forecasting. These models are characterized by their capabilities of learning long-term dependencies in collected electric data, which lead to accurate prediction results that outperform several alternative statistical and machine learning approaches. Unfortunately, applying LSTM models may not produce acceptable forecasting results, not only because of the noisy electric data but also due to the naive selection of its hyperparameter values. Therefore, an optimal configuration of an LSTM model is necessary to describe the electric consumption patterns and discover the time-series dynamics in the energy domain. Finding such an optimal configuration is, on the one hand, a combinatorial problem where selection is done from a very large space of choices; on the other hand, it is a learning problem where the hyperparameters should reflect the energy consumption domain knowledge, such as the influential time lags, seasonality, periodicity, and other temporal attributes. To handle this problem, we use in this paper metaheuristic-search-based algorithms, known by their ability to alleviate search complexity as well as their capacity to learn from the domain where they are applied, to find optimal or near-optimal values for the set of tunable LSTM hyperparameters in the electrical energy consumption domain. We tailor both a genetic algorithm (GA) and particle swarm optimization (PSO) to learn hyperparameters for load forecasting in the context of energy consumption of big data. The statistical analysis of the obtained result shows that the multi-sequence deep learning model tuned by the metaheuristic search algorithms provides more accurate results than the benchmark machine learning models and the LSTM model whose inputs and hyperparameters were established through limited experience and a discounted number of experimentations.

Efficient Hybrid Nature-Inspired Binary Optimizers for Feature Selection

Abstract: The process of dimensionality reduction is a crucial solution to deal with the dimensionality problem that may be faced when dealing with the majority of machine learning techniques. This paper proposes an enhanced hybrid metaheuristic approach using grey wolf optimizer (GWO) and whale optimization algorithm (WOA) to develop a wrapper-based feature selection method. The main objective of the proposed technique is to alleviate the drawbacks of both algorithms, including immature convergence and stagnation to local optima (LO). The hybridization is done with improvements in the mechanisms of both algorithms. To confirm the stability of the proposed approach, 18 well-known datasets are employed from the UCI repository. Furthermore, the classification accuracy, number of selected features, fitness values, and run time matrices are collected and compared with a set of well-known feature selection approaches in the literature. The results show the superiority of the proposed approach compared with both GWO and WOA. The results also show that the proposed hybrid technique outperforms other state-of-the-art approaches, significantly.

From ants to whales: metaheuristics for all tastes

Abstract: Nature-inspired metaheuristics comprise a compelling family of optimization techniques. These algorithms are designed with the idea of emulating some kind natural phenomena (such as the theory of evolution, the collective behavior of groups of animals, the laws of physics or the behavior and lifestyle of human beings) and applying them to solve complex problems. Nature-inspired methods have taken the area of mathematical optimization by storm. Only in the last few years, literature related to the development of this kind of techniques and their applications has experienced an unprecedented increase, with hundreds of new papers being published every single year. In this paper, we analyze some of the most popular nature-inspired optimization methods currently reported on the literature, while also discussing their applications for solving real-world problems and their impact on the current literature. Furthermore, we open discussion on several research gaps and areas of opportunity that are yet to be explored within this promising area of science.

Smart IoT Control-Based Nature Inspired Energy Efficient Routing Protocol for Flying Ad Hoc Network (FANET)

Abstract: The increasing demand of routing in the field of communication is the most important subject in ad hoc networks now a days. Flying Ad Hoc Network (FANET) is one of the emerging areas that evolved from Mobile Ad Hoc Networks. Selecting the best optimal path in any network is a real challenge for a routing protocol. Because the network performance like throughput, Quality of Service (QoS), user experience, response time and other key parameters depend upon the efficiency of the algorithm running inside the routing protocol. The complexity and diversity of the problem is augmented due to dynamic spatial and temporal mobility of FANET nodes. Due to these challenges the performance and efficiency of the routing protocol becomes very critical. This paper presents a novel routing protocol for FANET using modified AntHocNet. Ant colony optimization technique or metaheuristics in general has shown better dependability and performance as compared to other legacy best path selection techniques. Energy stabilizing parameter introduced in this study improves energy efficiency and overall network performance. Simulation results show that the proposed protocol is better than generic Ant Colony Optimization (ACO) and other traditional routing protocols utilized in FANET.

Search and Rescue optimization algorithm: a new optimization method for solving constrained engineering optimization problems

Abstract: A new optimization method namely the Search and Rescue optimization algorithm (SAR) is presented here to solve constrained engineering optimization problems. This metaheuristic algorithm imitates the explorations behavior of humans during search and rescue operations. The e-constrained method is utilized as a constraint-handling technique. Besides, a restart strategy is proposed to avoid local infeasible minima in some complex constrained optimization problems. SAR is applied to solve 18 benchmark constraint functions presented in CEC 2010, 13 benchmark constraint functions, and 7 constrained engineering design problems reported in the specialized literature. The performance of SAR is compared with some state-of-the-art optimization algorithms. According to the statistical comparison results, the performance of SAR is better or highly competitive against the compared algorithms on most of the studied problems.

Ant Lion Optimization: Variants, Hybrids, and Applications

Abstract: Ant Lion Optimizer (ALO) is a recent novel algorithm developed in the literature that simulates the foraging behavior of a Ant lions. Recently, it has been applied to a huge number of optimization problems. It has many advantages: easy, scalable, flexible, and have a great balance between exploration and exploitation. In this comprehensive study, many publications using ALO have been collected and summarized. Firstly, we introduce an introduction about ALO. Secondly, we categorized the recent versions of ALO into 3 Categories mainly Modified, Hybrid and Multi-Objective. we also introduce the applications in which ALO has been applied such as power, Machine Learning, Image processing problems, Civil Engineering, Medical, etc. The review paper is ended by giving a conclusion of the main ALO foundations and providing some suggestions & possible future directions that can be investigated.

Salp swarm algorithm: Theory, literature review, and application in extreme learning machines

Abstract: Salp Swarm Algorithm (SSA) is a recent metaheuristic inspired by the swarming behavior of salps in oceans. SSA has demonstrated its efficiency in various applications since its proposal. In this chapter, the algorithm, its operators, and some of the remarkable works that utilized this algorithm are presented. Moreover, the application of SSA in optimizing the Extreme Learning Machine (ELM) is investigated to improve its accuracy and overcome the shortcomings of its conventional training method. For verification, the algorithm is tested on 10 benchmark datasets and compared to two other well-known training methods. Comparison results show that SSA based training methods outperforms other methods in terms of accuracy and is very competitive in terms of prediction stability.