Deep-ELA: Deep Exploratory Landscape Analysis with Self-Supervised Pretrained Transformers for Single- and Multiobjective Continuous Optimization Problems.

Multi-objective optimization problems (MOP) are frequently encountered in practice. In some cases, different computationally expensive analyses may be independently used for computing different objectives of the MOP. Additionally, the analyses may be executed to obtain estimates with different fidelity, typically higher fidelity requiring a longer run-time. For instance, in automotive design, the aerodynamic drag is computed using computational fluid dynamic (CFD) analysis and its crashworthiness/strength is assessed using finite element analysis (FEA). Both the objectives can be independently computed and the underlying fidelity of each analysis can also be controlled using different mesh sizes/thresholds on the residual errors. While there exist a number of generic MOP benchmark problems in the literature, there is scarce work on constructing MOPs with multi-fidelity (MF) analyses to support the development of multi-fidelity, multi-objective optimization algorithms. The existing MF benchmarks are limited to unconstrained, single-objective optimization problems only. Towards addressing this gap, in this paper, we introduce a test suite for multi-objective, multi-fidelity optimization (MOMF). The problems are derived by combining existing unconstrained, multi-objective design optimization problems with resolution/stochastic/instability errors that are common manifestations of MF simulations. The method allows for the construction of any number of low-fidelity functions with desired level of correlations for a given high-fidelity objective function. We hope that the test suite would motivate novel algorithmic developments to support optimization involving computationally expensive and independently evaluable objectives.

Although various optimization methods for solving single-objective problems have been developed in the last few decades, these methods have lost their eligibility due to the fact that today’s engineering problems are toward multiple objective optimization problems, in real applications. For single-objective optimization problems, for example, in case of a minimization problem, this value is the decision vector giving the smallest objective that can be achieved within the specified constraints. Hence the minimum decision vector within all possible (feasible) solution vectors is the so-called optimal solution and/or optimal design. However, in multi-objective optimization problems, since a different objective value is generated against each decision vector, the superiority of the solutions over each other is determined by considering the trade-off among the objective values. Therefore, the solution of multi-objective optimization problems, unlike single-objective problems, is a set of vectors rather than a single decision vector. In multi-objective optimization problems, especially if there are intricate objectives, the computational cost of the problem increases. In other words, while synchronously trying to maximize one of the objectives and to minimize another one makes it difficult to find the global optimum design. One of the important techniques used in multi-objective optimization problems is Pareto optimality which enables to select the global optimum solution taking into account the trade-off among all objectives. In this context, using of derivative-based methods has decreased, but the use of metaheuristic methods has increased due to the rapid availability of global optimum solution. This is because the improvements in the field of optimization are progressing in proportion to technology and varying according to the needs. In this chapter, one of the recent metaheuristic optimization methods based on swarm intelligence that is so-called a Pareto-based bat algorithm inspired by the behavior of determining the direction and distance of an object using the echo of the sound called the echolocation of bats is used in order to obtain optimum solutions for multi-objective engineering design problems. In this regard, a four-bar planar truss, a real-sized welded steel beam as well as a multi-layer radar absorber are selected as multi-objective engineering design optimization problems. In case the obtained results (optimal designs) are examined, the potency and the reliability of the proposed multi-objective Pareto-based bat algorithm are demonstrated.

Construction site layout is multi-objective optimization (MOO) problem. However, the previous researchers treated site layout as a single objective optimization problem, which is far away from the reality. In this study, site layout was treated as MOO problem. The two objective functions are minimizing the likelihood of accidents happened and total handling cost of interaction flows between the facilities. The modified Pareto-based ant colony optimization (ACO) algorithm was proposed to solve site layout problem. Moreover, fuzzy rule-based system was used to quantify the interaction flows. Finally, a case of office building was used to evaluate the modified Pareto-based ACO algorithm for solving MOO site layout problem. The results showed the feasibility and effectiveness of the modified Pareto-based ACO algorithm to solve MOO site layout problem.

We propose a chaotic evolution algorithm with elite strategy. The conventional chaotic evolution algorithm uses each individual to search in its local area. The proposed algorithm searches the parameter space always around the elite individual from the last generation. We evaluate the proposed algorithm in both single-objective and multi-objective optimization problems. In the single objective optimization problem, the elite is the individual has the best fitness value, and in the multi-objective optimization problem, the elites are the individuals in the first Pareto front. We design and evaluate these two algorithms with elite strategy using single- and multi-objective benchmark functions. We design a jump strategy to avoid searching within a local optima areas by applying elite strategy several generations one time. The numerical evaluation results demonstrate the proposed algorithm has strong local exploitation capability in the early generations. The optimization performance of chaotic evolution algorithm has a potential possibility to apply in high dimensional and more complex optimization problems.

A multifactorial evolutionary algorithm (MFEA) is a recently proposed algorithm for evolutionary multitasking, which optimizes multiple optimization tasks simultaneously. With the design of knowledge transfer among different tasks, MFEA has demonstrated the capability to outperform its single-task counterpart in terms of both convergence speed and solution quality. In MFEA, the knowledge transfer across tasks is realized via the crossover between solutions that possess different skill factors. This crossover is thus essential to the performance of MFEA. However, we note that the present MFEA and most of its existing variants only employ a single crossover for knowledge transfer, and fix it throughout the evolutionary search process. As different crossover operators have a unique bias in generating offspring, the appropriate configuration of crossover for knowledge transfer in MFEA is necessary toward robust search performance, for solving different problems. Nevertheless, to the best of our knowledge, there is no effort being conducted on the adaptive configuration of crossovers in MFEA for knowledge transfer, and this article thus presents an attempt to fill this gap. In particular, here, we first investigate how different types of crossover affect the knowledge transfer in MFEA on both single-objective (SO) and multiobjective (MO) continuous optimization problems. Furthermore, toward robust and efficient multitask optimization performance, we propose a new MFEA with adaptive knowledge transfer (MFEA-AKT), in which the crossover operator employed for knowledge transfer is self-adapted based on the information collected along the evolutionary search process. To verify the effectiveness of the proposed method, comprehensive empirical studies on both SO and MO multitask benchmarks have been conducted. The experimental results show that the proposed MFEA-AKT is able to identify the appropriate knowledge transfer crossover for different optimization problems and even at different optimization stages along the search, which thus leads to superior or competitive performances when compared to the MFEAs with fixed knowledge transfer crossover operators.

A Multiobjective Optimization Problem (MOP) requires the optimization of several objective functions simultaneously, usually in conflict with each other. One of the most efficient algorithms for solving MOPs is MOEA/D (Multiobjective Evolutionary Algorithm Based on Decomposition), which decomposes a MOP into single-objective optimization subproblems and solves them using information from neighboring subproblems. MOEA/D variants with other evolutionary operators have emerged over the years, improving their efficiency in various MOPs. Recently, the IWOA (Improved Whale Optimization Algorithm) was proposed, an optimization algorithm bioinspired by the whale hunting method hybridized with Differential Evolution, which presented excellent results in single-objective optimization problems. This work proposes the MOEA/D-IWOA algorithm, which associates characteristics of the evolutionary operators of the IWOA to MOEA/D. Computational experiments were accomplished to analyze the performance of the MOEA/D-IWOA in benchmark MOPs suites. The results were compared with those obtained by the MOEA/D, Non-dominated Sorting Genetic Algorithm II (NSGA-II), Third Evolution Step of Generalized Differential Evolution (GDE3), Improving the Strength Pareto Evolutionary Algorithm (SPEA2), and Indicator-Based Evolutionary Algorithm (IBEA) algorithms in the Hypervolume and Inverted Generational Distance Plus (IGD+) indicators. The MOEA/D-IWOA proved to be competitive, with a good performance profile, in addition to presenting the best results in some POMs.

Multi-objective optimization problems can be divided into continuous multi-objective optimization problems and discrete multi-objective optimization problems, and discrete multi-objective optimization is not universal. In practical applications, there are many discrete multi-objective optimization problems. The solution of different problems needs to design different evolutionary multi-objective algorithms according to their specific conditions. The threshold in the traditional multi-objective optimization of wireless network is a preset constant. It has relatively poor performance in the synthesis of images with different noise levels. An adaptive wireless network multi-objective optimization algorithm based on image synthesis is proposed. Based on the maximum inter class variance and maximum peak signal to noise ratio (SNR), an adaptive wireless network multi-objective optimization algorithm is established. The accuracy and noise immunity of image synthesis are also considered. In order to avoid the effect of threshold increase on algorithm efficiency, multi-objective optimization algorithm is introduced into the algorithm. Experiments show that the method proposed in this paper is accurate and robust and has good universality for the synthesis of different noise images.

Runtime analysis of a multi-objective evolutionary algorithm for obtaining finite approximations of Pareto fronts

Multiobjective Optimization Problems (MOP) can be found in many issues of scientific research, engineering, and in everyday social life. A MOP problem has several objectives that conflict with one another which must be optimized simultaneously. This paper presents a quantum-inspired evolutionary algorithm (QEA) to solve continuous multiobjective optimization problem (MOP). The proposed method employs Fast Nondominated Sorting and Crowding Distance from NSGA-II and implements all common operators of genetic algorithms (GA), such as crossover and mutations with additional Quantum Gate quantum operators. The proposed method is then run in a distributed manner and is proven to be able to significantly outperform the hypervolume and MOEA/D metrics and have hypervolumes that are comparable to NSGA-II while maintaining a better average Δ’ in all testing problems. From this result, it is concluded that using quantum-inspired individual genetic algorithms to solve continuous MOP can produce hypervolume and Δ’ metrics that are good in all specified test problems.

In this paper, we propose a novel multi-objective ant colony optimizer (called iMOACO\(_{\mathbb {R}}\)) for continuous search spaces, which is based on ACO\(_{\mathbb {R}}\) and the R2 performance indicator. iMOACO\(_{\mathbb {R}}\) is the first multi-objective ant colony optimizer (MOACO) specifically designed to tackle continuous many-objective optimization problems (i.e., multi-objective optimization problems having four or more objectives). Our proposed iMOACO\(_{\mathbb {R}}\) is compared to three state-of-the-art multi-objective evolutionary algorithms (NSGA-III, MOEA/D and SMS-EMOA) and a MOACO algorithm called MOACO\(_{\mathbb {R}}\) using standard test problems and performance indicators taken from the specialized literature. Our experimental results indicate that iMOACO\(_{\mathbb {R}}\) is very competitive with respect to NSGA-III and MOEA/D and it is able to outperform SMS-EMOA and MOACO\(_{\mathbb {R}}\) in most of the test problems adopted.

This paper deals with the multi-objective optimization of tire reinforcement structures such as the tread belt and the carcass path. The multi-objective functions are defined in terms of the discrete-type design variables and approximated by artificial neutral network, and the sensitivity analyses of these functions are replaced with the iterative genetic evolution. The multi-objective optimization algorithm introduced in this paper is not only highly CPU-time-efficient but it can also be applicable to other multi-objective optimization problems in which the objective function, the design variables and the constraints are not continuous but discrete. Through the illustrative numerical experiments, the fiber-reinforced tire belt structure is optimally tailored. The proposed multi-objective optimization algorithm is not limited to the tire reinforcement structure, but it can be applicable to the generalized multi-objective structural optimization problems in various engineering applications.

Recently, the particle swarm algorithm (PSO) has demonstrated its effectiveness in solving multi-objective optimization problems. However, the performance of most existing multi-objective particle swarm algorithms depends largely on the global or individual best particles. Moreover, due to the rapid convergence of PSO in single objective optimization problems, PSO is prone to poorly distributed indicators when dealing with multi-objective optimization problems. To solve the above problems, we propose a multi-objective competitive particle swarm algorithm based on vector angles (VaCSO). Firstly, in order to remove the influence of global best particles or individual best particles on the algorithm, the competition mechanism is used. Secondly, in order to increase the diversity of solutions while maintaining the convergence of the algorithm, the population is clustered into two populations. Population 1 mainly considers the convergence of the solution in the offspring generation strategy. As a supplement, population 2 adds a new offspring generation strategy to maintain the distribution of the solution, and we innovatively proposed a three-particle competition to improve the distribution and diversity of particle swarms. Finally, based on vector angle information, we consider auxiliary learning to optimize the population gap, so as to improve the distribution of the algorithm. We have established two sets of comparative experiments to test the performance of VaCSO. We compared VaCSO with the currently popular multi-objective particle swarm optimizers and multi-objective evolutionary algorithms. Experimental results show that VaCSO has an excellent performance in convergence and distribution, and has a significant effect in optimizing quality.

Decision makers tend to define their optimization problems as multi-objective optimization problems. Generating the whole nondominated set is often unrealistic due to its size but also because most of these points are irrelevant to the decision maker. Another approach consists in obtaining preference information and integrating it a priori, in order to reduce the size of the nondominated set and have a gain in computation time. In this work we focus on a partial preference relation based on requirement and tolerance thresholds that translate the Pareto dominance cone. After introducing this preference relation, we present adaptations to use it in existing discrete multi-objective optimization algorithms. Numerical experiments on multi-objective combinatorial optimization problems show the applicabiliy of our approach.

In a real life situation, due to the complexity of the industrial systems and non-linearity of their behavior, it is very difficult to achieve optimum performance of system for desired industrial goals using uncertain, vague and imprecise data. Herein, an approach has been proposed through which the behavior of the system is analyzed in the form of well- known six reliability indices by using triangular fuzzy numbers which allow the consideration of expert opinions, linguistic variables, operating conditions in reliability information. Using their behavior analysis a fuzzy multi-objective optimization problem (FMOOP) has been formulated. Due to the conflicting nature of the multiple objectives, the decision making is difficult and it leads to the Pareto optimal solutions instead of single optimal solutions. Many evolutionary algorithms (EAs) already exist in the literature for solving a multi-objective optimization problem (MOOP), and are termed as multi-objective evolutionary algorithms (MOEAs). Particle swarm optimization (PSO) is one of such MOEA which demonstrates the ability to identity a Pareto-optimal front efficiently. Here, a crisp optimization problem is reformulated from FMOOP by taking into account the preference of decision maker (DM) and then PSO is applied to solve the resulting fuzzified MOOP. The presented approach is applied in order to solve the multi-objective series-parallel system reliability optimization problem for a crystallization unit of a urea plant.

When selecting the best suited algorithm for an unknown optimization problem, it is useful to possess some a priori knowledge of the problem at hand. In the context of single-objective, continuous optimization problems such knowledge can be retrieved by means of Exploratory Landscape Analysis (ELA), which automatically identifies properties of a landscape, e.g., the so-called funnel structures, based on an initial sample. In this paper, we extract the relevant features (for detecting funnels) out of a large set of landscape features when only given a small initial sample consisting of 50 x D observations, where D is the number of decision space dimensions. This is already in the range of the start population sizes of many evolutionary algorithms. The new Multiple Peaks Model Generator (MPM2) is used for training the classifier, and the approach is then very successfully validated on the Black-Box Optimization Benchmark (BBOB) and a subset of the CEC 2013 niching competition problems.