Search

The integration of Machine Learning (ML) and Constrained Optimization (CO) techniques has recently gained significant interest. While pure CO methods struggle with scalability and robustness, and ML methods like constrained Reinforcement Learning (RL) face difficulties with combinatorial decision spaces and hard constraints, a hybrid approach shows promise. However, multi-stage decision-making under uncertainty remains challenging for current methods, which often rely on restrictive assumptions or specialized algorithms. This paper introduces unify , a versatile framework for tackling a wide range of problems, including multi-stage decision-making under uncertainty, using standard ML and CO components. unify integrates a CO problem with an unconstrained ML model through parameters controlled by the ML model, guiding the decision process. This ensures feasible decisions, minimal costs over time, and robustness to uncertainty. In the empirical evaluation, unify demonstrates its capability to address problems typically handled by Decision Focused Learning, Constrained RL, and Stochastic Optimization. While not always outperforming specialized methods, unify 's flexibility offers broader applicability and maintainability. The paper includes the method's formalization and empirical evaluation through case studies in energy management and production scheduling, concluding with future research directions. [ABSTRACT FROM AUTHOR]

Abstract This paper proposes a novel optimization algorithm, called Emperor Penguin Optimizer (EPO), which mimics the huddling behavior of emperor penguins (Aptenodytes forsteri). The main steps of EPO are to generate the huddle boundary, compute temperature around the huddle, calculate the distance, and find the effective mover. These steps are mathematically modeled and implemented on 44 well-known benchmark test functions. It is compared with eight state-of-the-art optimization algorithms. The paper also considers for solving six real-life constrained and one unconstrained engineering design problems. The convergence and computational complexity are also analyzed to ensure the applicability of proposed algorithm. The experimental results show that the proposed algorithm is able to provide better results as compared to the other well-known metaheuristic algorithms. [ABSTRACT FROM AUTHOR]

In this paper a novel nature-inspired optimization paradigm is proposed called Moth-Flame Optimization (MFO) algorithm. The main inspiration of this optimizer is the navigation method of moths in nature called transverse orientation. Moths fly in night by maintaining a fixed angle with respect to the moon, a very effective mechanism for travelling in a straight line for long distances. However, these fancy insects are trapped in a useless/deadly spiral path around artificial lights. This paper mathematically models this behaviour to perform optimization. The MFO algorithm is compared with other well-known nature-inspired algorithms on 29 benchmark and 7 real engineering problems. The statistical results on the benchmark functions show that this algorithm is able to provide very promising and competitive results. Additionally, the results of the real problems demonstrate the merits of this algorithm in solving challenging problems with constrained and unknown search spaces. The paper also considers the application of the proposed algorithm in the field of marine propeller design to further investigate its effectiveness in practice. Note that the source codes of the MFO algorithm are publicly available at http://www.alimirjalili.com/MFO.html . [ABSTRACT FROM AUTHOR]

This paper proposes a multi-objective version of recently developed Spotted Hyena Optimizer (SHO) called Multi-objective Spotted Hyena Optimizer (MOSHO). It is used to optimize the multiple objectives problems. In the proposed algorithm, a fixed-sized archive is employed for storing the non-dominated Pareto optimal solutions. The roulette wheel mechanism is used to select the effective solutions from archive to simulate the social and hunting behaviors of spotted hyenas. The proposed algorithm is tested on 24 benchmark test functions and compared with six recently developed metaheuristic algorithms. The proposed algorithm is then applied on six constrained engineering design problems to demonstrate its applicability on real-life problems. The experimental results reveal that the proposed algorithm performs better than the others and produces the Pareto optimal solutions with high convergence. [ABSTRACT FROM AUTHOR]

Gravitational search algorithm (GSA), a recent meta-heuristic algorithm inspired by Newton's law of gravity and mass interactions, shows good performance in various optimization problems. In GSA, the gravitational constant attenuation factor alpha ( α ) plays a vital role in convergence and the balance between exploration and exploitation. However, in GSA and most of its variants, all agents share the same α value without considering their evolutionary states, which has inevitably caused the premature convergence and imbalance of exploration and exploitation. In order to alleviate these drawbacks, in this paper, we propose a new variant of GSA, namely stability constrained adaptive alpha for GSA (SCAA). In SCAA, each agent's evolutionary state is estimated, which is then combined with the variation of the agent's position and fitness feedback to adaptively adjust the value of α . Moreover, to preserve agents’ stable trajectories and improve convergence precision, a boundary constraint is derived from the stability conditions of GSA to restrict the value of α in each iteration. The performance of SCAA has been evaluated by comparing with the original GSA and four alpha adjusting algorithms on 13 conventional functions and 15 complex CEC2015 functions. The experimental results have demonstrated that SCAA has significantly better searching performance than its peers do. [ABSTRACT FROM AUTHOR]

Given a task requiring a set of skills, the team formation problem in social network aims to find a team that covers all the required skills and minimizes the communication cost. However, to make the team work more efficiently, a proper leader is needed for managing and communicating with all the members in the team. As the number of required skills increases and the team grows in size, a single leader is not sufficiently capable of administering a large project since the leader may not have enough time to communicate with all of the team members. Therefore, from the practical perspective, the team would be divided into smaller sub-teams, each of them has a leader. In the field of management, the size constraint of each team is called Span of Control (SoC). In this paper, we tackle the problem of finding teams of experts with SoC constraints in social network. To solve the problem effectively, we propose three basic algorithms, and explore a Self-Organizing mechanism to determine the role of each individual in a team to be the leader or a member. For solving the problem more efficiently, we design a γ HSCCent node ranking strategy, and the corresponding enhanced algorithms are proposed. Experimental results from both quantitative and qualitative studies show that the teams composed by our proposed algorithms have better performance in both effectiveness and efficiency. [ABSTRACT FROM AUTHOR]

Constrained optimization (CO) has made a profound impact in solving many real-world problems. Due to the high computation burden in exact solvers, data-driven CO based on machine learning techniques is recently receiving extensive research interests for its capability to solve CO problems in real time. The existing data-driven CO approaches only serve for optimization problems with rather simple constraints that can be directly incorporated into model training. However, constraints that are computationally infeasible or burdensome to evaluate are commonly experienced in realistic optimization applications, especially in the engineering sector. This paper proposes an athlete–referee dual learning system (ARDLS) for end-to-end CO with large-scale complex constraints, where an athlete model is trained as the main optimizer while a referee model is trained as a probabilistic constraint classifier to guide the athlete training. A risk-based constrained loss function is designed to fine-tune the athlete model for constraint satisfaction. A case study on electric power system emergency control application is conducted to validate the proposed ARDLS, where the testing results demonstrate the excellent capability of ARDLS to improve the likelihood of satisfying large-scale complex constraints in CO. [ABSTRACT FROM AUTHOR]

Design optimization of passive microwave components is an intricate process, especially if the primary objective is a reduction of the physical size of the structure. The latter has become an important design consideration for a growing number of modern applications (mobile communications, wearable/implantable devices, internet of things), where miniaturization is imperative due to a limited space allocated for the electronic circuitry. Optimization-based size reduction is a heavily constrained task, with several acceptance thresholds imposed on electrical characteristics of the system. The challenges are pronounced whenever equality constraints are involved (e.g., related to power split ratio requirements), in which case the feasible space is a thin set, thereby difficult to be explored throughout the optimization process. This feature makes conventional methods, such as penalty function approaches or algorithms with explicit constraint handling, of limited reliability. In this paper, we introduce a novel technique for reliable control of equality constraints in simulation-driven size reduction of microwave components. Our methodology involves an intermittent optimization-based correction of equality constraints. This is essentially a knowledge-based decision-making strategy implemented as a supplementary optimization stage, and launched before each iteration of the core algorithm. Constraint violation is reduced without being detrimental to the remaining figures of merit, in particular, the circuit size and inequality constraints. Meanwhile, inequality constraints are handled using a penalty function approach with adaptive adjustment of penalty coefficients. The proposed technique facilitates exploration of the feasible space, and allows for achieving reduced miniaturization rates in comparison to the benchmark methods, while ensuring a reliable control of the design constraints. These advantages have been demonstrated using four microstrip couplers, with consistent results obtained for all considered circuits. [ABSTRACT FROM AUTHOR]

The single objective quadratic multiple knapsack problem (QMKP) is a useful model to formulate a number of practical problems. However, it is not suitable for situations where more than one objective needs to be considered. In this paper, we extend the single objective QMKP to the bi-objective case such that we simultaneously maximize the total profit of the items packed into the knapsacks and the ’makespan’ (the gain of the least profit knapsack). Given the imposing computational challenge, we propose a hybrid two-stage (HTS) algorithm to approximate the Pareto front of the bi-objective QMKP. HTS combines two different and complementary search methods — scalarizing memetic search (first stage) and Pareto local search (second stage). Experimental assessments on a set of 60 problem instances show that HTS dominates a standard multi-objective evolutionary algorithm (NSGA II), and two simplified variants of HTS. We also present a comparison with two state-of-the-art algorithms for the single objective QMKP to assess the quality of the extreme solutions of the approximated Pareto front. [ABSTRACT FROM AUTHOR]

This paper proposes an entropy-based bare bones particle swarm for solving dynamic constrained optimization problems. The Shannon’s entropy is established as a phenotypic diversity index and the proposed algorithm uses the Shannon’s index of diversity to aggregate the global-best and local-best bare bones particle swarm variants. The proposed approach applies the idea of mixture of search directions, using the index of diversity as a factor to balance the influence of the global-best and local-best search directions. High diversity promotes the search guided by the global-best solution, with a normal distribution for exploitation. Low diversity promotes the search guided by the local-best solution, with a heavy-tailed distribution for exploration. A constraint-handling strategy is also proposed, which uses a ranking method with selection based on the technique for order preference by similarity to ideal solution (TOPSIS) to obtain the best solution within a specific population of candidate solutions. Mechanisms to detect changes in the environment and to update particles’ memories are also implemented into the proposed algorithm. All these strategies do not act independently. They operate related to each other to tackle problems such as: diversity loss due to convergence and outdated memories due to changes in the environment. The combined effect of these strategies provides an algorithm with ability to maintain a proper balance between exploration and exploitation at any stage of the search process without losing the tracking ability. An empirical study was carried out to evaluate the performance of the proposed approach. Experimental results show the suitability of the algorithm in terms of effectiveness to find good solutions for the benchmark problems investigated. Finally, an application is developed where the proposed algorithm is applied to solve the dynamic economic dispatch problem in power systems. [ABSTRACT FROM AUTHOR]

This paper proposes a novel population-based optimization algorithm called Sine Cosine Algorithm (SCA) for solving optimization problems. The SCA creates multiple initial random candidate solutions and requires them to fluctuate outwards or towards the best solution using a mathematical model based on sine and cosine functions. Several random and adaptive variables also are integrated to this algorithm to emphasize exploration and exploitation of the search space in different milestones of optimization. The performance of SCA is benchmarked in three test phases. Firstly, a set of well-known test cases including unimodal, multi-modal, and composite functions are employed to test exploration, exploitation, local optima avoidance, and convergence of SCA. Secondly, several performance metrics (search history, trajectory, average fitness of solutions, and the best solution during optimization) are used to qualitatively observe and confirm the performance of SCA on shifted two-dimensional test functions. Finally, the cross-section of an aircraft's wing is optimized by SCA as a real challenging case study to verify and demonstrate the performance of this algorithm in practice. The results of test functions and performance metrics prove that the algorithm proposed is able to explore different regions of a search space, avoid local optima, converge towards the global optimum, and exploit promising regions of a search space during optimization effectively. The SCA algorithm obtains a smooth shape for the airfoil with a very low drag, which demonstrates that this algorithm can highly be effective in solving real problems with constrained and unknown search spaces. Note that the source codes of the SCA algorithm are publicly available at http://www.alimirjalili.com/SCA.html . [ABSTRACT FROM AUTHOR]

The quadratic multiple knapsack problem (QMKP) is a challenging combinatorial optimization problem with numerous applications. In this paper, we propose the first evolutionary path relinking approach (EPR) for solving the QMKP approximately. This approach combines advanced features both from the path relinking (PR) method and the responsive threshold search algorithm. Thanks to the tunneling property which allows a controlled exploration of infeasible regions, the proposed EPR algorithm is able to identify very high quality solutions. Experimental studies on the set of 60 well-known benchmarks and a new set of 30 large-sized instances show that EPR outperforms several state-of-the-art algorithms. In particular, for the 60 conventional benchmarks, it discovers 10 improved results (new lower bounds) and matches the best known result for the remaining 50 cases. More significantly, EPR demonstrates remarkable efficacy on the 30 new larger instances by easily dominating the current best performing algorithms across the whole instance set. Key components of the algorithm are analyzed to shed lights on their impact on the proposed approach. [ABSTRACT FROM AUTHOR]

In recent years, the uptake of Artificial Intelligence (AI) in industry is increasing. For many AI techniques, like Deep Learning, optimization, planning, etc., computational and storage requirements are significant. The problem of determining what is the right hardware (HW on premise or on the cloud) architecture and its dimensioning for AI algorithms is still crucial. Searching for the optimal solution is often challenging, as it is not trivial to anticipate the behavior of an algorithm on diverse architectures. This is especially true if the AI application must respect quality-of-service constraints or budgets. In this scenario, having an automated decision support tool to match algorithms, user constraints and HW resources would be a great advantage for companies and practitioners working with AI applications. In this paper, we tackle this challenge with an approach that relies on the Empirical Model Learning paradigm, based on the integration of Machine Learning (ML) models into an optimization problem. The key idea is to integrate domain knowledge held by experts with data-driven models that learn the relationships between HW requirements and AI algorithm performances. In particular, the approach starts with benchmarking multiple AI algorithms on different HW resources, generating data used to train ML models; then, optimization is used to find the best HW configuration that respects user-defined constraints (e.g., budget, time, solution quality). In the experimental evaluation we validate our approach on a complex problem, namely online algorithms for energy systems, an area characterized by uncertainty and tight HW and real-time constraints. Results show the effectiveness of our approach and its flexibility: We can train the ML models only once and reuse them in the optimization model to tackle a variety of problems, determined by different data instances and user-defined constraints. [ABSTRACT FROM AUTHOR]

When solving constrained multiobjective optimization problems, it is crucial to deal with several conflicting objectives and various constraints simultaneously. To address these two issues, a constrained multiobjective differential evolution algorithm with an infeasible-proportion control mechanism is presented in this paper. Specifically, two populations are cooperatively employed in the evolution process. The first population is used to solve the original problem, that is, to explore the constrained Pareto front, while the second population is created to search for high-quality objective function information. Furthermore, differential evolution with two specific mutation strategies is employed to update each population. Cooperative strategies can not only balance diversity and convergence but also realize information exchange between the two populations provide new evolutionary directions. In addition, an infeasible-proportion control mechanism is used to gradually decrease the proportion of infeasible solutions in the second population. This enables the main population to pass through the infeasible region barrier in the early evolution stage and facilitates the search for a constrained Pareto front in the later evolution stage. Systematic experiments on 65 benchmark test functions show that the proposed algorithm is superior to or at least comparable to five well-established constrained multiobjective optimization methods. [ABSTRACT FROM AUTHOR]

Solving constrained multi-objective optimization problems (CMOPs) obtains considerable attention in the evolutionary computation community. Various constrained multi-objective evolutionary algorithms (CMOEAs) have been developed for the CMOPs in the last few decades. Among the CMOEA techniques, two archive strategy is an effective approach, and enhancing the performance of C-TAEA based on two archive framework is a promising direction. This paper proposes an improved two-archive-based evolutionary algorithm, referred to as C-TAEA2. In C-TAEA2, a new fitness evaluation strategy for the convergence archive (CA) is presented to achieve better convergence. Additionally, a fitness evaluation method is proposed to evaluate solutions of the diversity archive (DA) to further promote diversity. Moreover, new update strategies are designed for both CA and DA to reduce the computational cost. Based on the new fitness evaluation strategies, a new mating selection strategy is also developed. Experiments on different benchmark CMOPs demonstrate that C-TAEA2 obtained better or highly competitive performance compared to other state-of-the-art CMOEAs. • An improved two-archive based EA is proposed for the CMOPs. • New fitness evaluation strategies are designed for the two archives. • An enhanced mating selection strategy is developed. • Results demonstrate the superiority of our approach to other CMOEAs. [ABSTRACT FROM AUTHOR]

Distributed particle filters (PFs) have received extensive attention because of its excellent performance. Generally, it needs to transmit particles to multiple processing units (PUs) in order to improve the performance. However, how to balance communication costs and computation costs in particle routing is still an open problem. This paper presents a multi-objective discrete particle swarm optimization (PSO) algorithm to solve the problem. In the algorithm, the particle routing problem in distributed PFs is modeled as a multi-objective constrained optimization model for the first time. Following that, an improved hybrid discrete multi-objective PSO is proposed. Two new operators designed based on the problem's characteristics, that is, a local search strategy based on molecular force and a constraint processing mechanism with greedy search, are developed to improve the performance of the proposed algorithm. By comparing with three commonly used methods on classical particle filters problems, experimental results show that the proposed algorithm is a highly competitive approach, and it can provide multiple high-quality Pareto optimal solutions for decision-makers to meet their different needs. [Display omitted] [ABSTRACT FROM AUTHOR]

Constrained multi-objective evolutionary algorithms (CMOEAs) have been extensively studied in recent years. However, the performance of most of traditional CMOEAs is unsatisfied for constrained multi-objective optimization problems (CMOPs) with small feasible regions. Based on the idea of two-archive, this paper proposes a novel two-archive evolutionary algorithm for constrained multi-objective optimization with small feasible regions. Specifically, we maintain two archives, named convergence-oriented archive (CA) and diversity-oriented archive (DA). To handle the CMOPs which feasible regions are small and far from the unconstrained Pareto front (PF), a cooperation-based mating selection mechanism is proposed. To strike a balance among convergence, diversity, and feasibility, a high-quality solution selection mechanism is proposed, which can help the CA approach PF from different directions and balance the convergence and diversity. To provide better diversity, a dynamic selection strategy is designed to update DA according to the status of the CA. In addition, in order to make the population evenly distributed in feasible regions, a replacement mechanism of the ideal point is designed. Compared with the four state-of-the-art constrained multi-objective evolutionary optimization algorithms, comprehensive experiments on a series of benchmark problems fully demonstrate the superiority of the proposed algorithm in terms of Δ p and hypervolume (HV). [ABSTRACT FROM AUTHOR]

When using evolutionary algorithms to address constrained optimization problems, it is important to balance not only the diversity and convergence but also the constraints and objective function. To this end, a self-adaptive resources allocation-based differential evolution (SRADE) is presented in this paper. Specifically, during the evolutionary process, three mutation strategies with distinct focuses are collaboratively employed and adaptively assigned to different individuals based on their performance feedback. That is, most of the computing resources will be consumed by the most efficient strategy at different evolutionary stages to mitigate inefficient search under limited resources. These three collaborative strategies focus on maintaining population diversity, driving the population into feasible regions, and promoting the population toward the objective, respectively. Combining the self-adaptive resources allocation scheme and diverse search strategies is expected to satisfy the requirements of the population for diversity, convergence, constraints, and the objective at a certain iteration. Extensive experiments are performed on three benchmark test suites, including a large number of test functions from IEEE CEC 2006, 2010, and 2017. Compared to other well-designed constrained evolutionary approaches, SRADE exhibits superior or very competitive performance. [ABSTRACT FROM AUTHOR]

Wind speed interval forecasting is being considered increasingly because it can provide more comprehensive information to address uncertainties in wind power generation. It helps ensure power quality, optimize power dispatching, and bring more economic and social benefits. Currently, the lower upper bound estimation (LUBE) approach is believed to provide outstanding performance in interval forecasting. However, considering the significant noise and irregular characteristics of wind speed, most existing LUBE models can either ineffectively learn the wind speed variation patterns behind the data or behave in a very unstable manner, leading to unsatisfactory results. Thus, this paper proposes a novel LUBE-based wind speed interval forecasting system based on an innovative parallel feature selection module for extracting the underlying vital historical variation patterns and a unique constrained LUBE training algorithm characterized by an amnesia operator to further guarantee the efficiency and stability of the LUBE training. The system effectiveness was demonstrated by performing experiments using two real datasets. The results show that the proposed system performs better than the naive, bootstrap, error analysis, and other LUBE models. It at least enhances the coverage width criterion by 1.8% and 6.8% for the two datasets, respectively. • The LUBE approach is developed into a systematic framework for enhancement. • A novel parallel feature selection module is proposed to extract vital factors. • The original GOA is modified to address the constrained optimization problem. • A so-called amnesia operator is proposed to improve the LUBE training process. • The proposed system enhances the quality of wind speed interval forecasting. [ABSTRACT FROM AUTHOR]

Abstract: The increasing complexity of circuit design needs to be managed with appropriate optimization algorithms and accurate statistical description of design models in order to reach the design specifics, guaranteeing ”zero defects”. In the Design for Yield open problems are the design of effective optimization algorithms and statistical analysis for yield design, which require time consuming techniques. New methods have to balance accuracy, robustness and computational effort. Typical analog integrated circuit optimization problems are computationally hard and require the handling of multiple, conflicting, and non-commensurate objectives having strong nonlinear interdependence. This paper tackles the problem by evolutionary algorithms to produce tradeoff solutions on the Pareto Front. In this research work Integrated Circuit (IC) design has been formulated as a constrained multi-objective optimization problem defined in a mixed integer/discrete/continuous domain. The following real-life circuits, RF Low Noise Amplifier, LeapFrog Filter, and Ultra Wideband LNA, were selected as test bed. The proposed algorithm, A-NSGAII, was shown to produce acceptable and robust solutions in the tested applications, where state-of-art algorithms and circuit designers failed. The results show significant improvement in all the chosen IC design problems. [Copyright &y& Elsevier]

The widespread existence of constrained multi-objective optimization problems (CMOPs) in practical applications encourages researchers to devote more efforts to the development of constrained multi-objective evolutionary algorithms (CMOEAs). An efficient way of handling CMOPs is to balance the satisfaction of constrains and the optimization of objective functions. Although many approaches have been designed to this end, most of them are still unable to handle CMOPs with diverse characteristics effectively. To remedy this issue, a simple and generic two-stage framework is proposed in this paper to achieve better efficiency and versatility for CMOPs. Specifically, the whole search process is simply divided into two stages with different purposes: (i) Stage 1 focuses on approaching the unconstrained Pareto front (PF) and storing the obtained feasible solutions in the archive. During stage 1 , the archive is updated to promote convergence and diversity; (ii) Stage 2 is mainly designed to obtain the constrained PF of CMOPs. First, the solutions in the archive are used to form the initial population. Afterwards, the constrained Pareto front (CPF) is extensively explored. Based on this framework, a CMOEA, referred to as C-TSEA, is presented. Fifty-seven instances in five benchmark test suites are chosen to evaluate the performance of our approach. Compared with seven advanced CMOEAs, the results demonstrate the superiority or at least competitiveness of C-TSEA. The generality of the proposed framework is also verified by integrating other methods in the proposed framework. • A simple and generic two-stage framework is proposed for CMOPs. • C-TSEA is presented based on the framework to effectively solve CMOPs. • Fifty-seven CMOPs are chosen to extensively evaluate its performance. • C-TSEA gets better or competitive results compared with other methods. [ABSTRACT FROM AUTHOR]

Recently, decomposition has gained a wide interest in solving multi-objective optimization problems involving more than three objectives also known as Many-objective Optimization Problems (MaOPs). In the last few years, there have been many proposals to use decomposition to solve unconstrained problems. However, fewer is the amount of works that has been devoted to propose new decomposition-based algorithms to solve constrained many-objective problems. In this paper, we propose the ISC-Pareto dominance (Isolated Solution-based Constrained Pareto dominance) relation that has the ability to: (1) handle constrained many-objective problems characterized by different types of difficulties and (2) favor the selection of not only infeasible solutions associated to isolated sub-regions but also infeasible solutions with smaller CV (Constraint Violation) values. Our constraint handling strategy has been integrated into the framework of the Constrained Non-Dominated Sorting Genetic Algorithm-III (C-NSGA-III) to produce a new algorithm called Isolated Solution-based Constrained NSGA-III (ISC-NSGA-III). The empirical results have demonstrated that our constraint handling strategy is able to provide better and competitive results when compared against three recently proposed constrained decomposition-based many-objective evolutionary algorithms in addition to a penalty-based version of NSGA-III on the CDTLZ benchmark problems involving up to fifteen objectives. Moreover, the efficacy of ISC-NSGA-III on a real world water management problem is showcased. [ABSTRACT FROM AUTHOR]

In this paper, we study on how to achieve balance between minimizing the objectives, satisfying the constraints, and avoiding the population to stuck at locally optimal or locally feasible regions. We propose a new constraints handling technique based on diversity distance measure for constrained multiobjective optimization problems. We innovatively measure all effects of the constraints on the objectives by the interaction effects and direct effects of constraints on the objectives. The interactive effects of constraints on the objective functions are firstly expressed and quantified by using Mahalanobis distance and computing the value of the distance measure by involving a Lebesgue measure. Based on interactive and direct effects, we propose a new modified mechanism of objective space, called objective space modified mechanism based on diversity distance measure, so that the good infeasible solution could be more effectively used to find the optimal solution. The Pull and Push search is used to adaptively adjust the location of the Pareto front, which could preclude the population from being stuck at some locally optimal or locally feasible regions and could decrease time complexity. Comprehensive experiments completed for several benchmark problems demonstrate the competitiveness of the proposed algorithm, in comparison to the existing state-of-art constrained evolutionary multiobjective optimization. [ABSTRACT FROM AUTHOR]

Surrogate-assisted evolutionary algorithms have been commonly used in extremely expensive optimization problems. However, many existing algorithms are only significantly used in continuous and unconstrained optimization problems despite the fact that plenty of real-world problems are constrained combinatorial optimization problems. Therefore, a random forest assisted adaptive multi-objective particle swarm optimization (RFMOPSO) algorithm is proposed in this paper to address this challenge. Firstly, the multi-objective particle swarm optimization (MOPSO) combines with random forest model to accelerate the overall search speed of the algorithm. Secondly, an adaptive stochastic ranking strategy is performed to balance better objectives and feasible solutions. Finally, a novel rule is developed to adaptively update the states of particles. In order to validate the proposed algorithm, it is tested by ten multi-objective knapsack benchmark problems whose discrete variables vary from 10 to 100. Experimental results demonstrate that the proposed algorithm is promising for optimizing the constrained combinatorial optimization problem. • A surrogate-assisted multi-objective particle swarm optimization algorithm is proposed. • Using adaptive stochastic ranking strategy to improve the quality of parent individuals. • A novel state updating mechanism is developed to guide the searching of particles. [ABSTRACT FROM AUTHOR]

Machine learning models are very sensitive to the hyperparameters, and their evaluations are generally expensive. Users desperately need intelligent methods to quickly optimize hyperparameter settings according to known evaluation information, so as to effectively promote the performance of the machine learning models within the limited and small budget. Motivated by this, in this paper, we propose ExperienceThinking algorithm to quickly find the best possible hyperparameter configuration of machine learning algorithms within a few configuration evaluations. ExperienceThinking designs two novel approaches, which make full use of the known evaluation information to intelligently infer optimal configurations from two aspects: search space pruning and knowledge utilization respectively. Two approaches suit for two different kinds of constrained hyperparameter optimization problems, they complement with each other and their combination increases the generality and effectiveness of the ExperienceThinking. To demonstrate the benefit of ExperienceThinking, we conduct extensive experiments using various constrained hyperparameter optimization problems, and compare it with classic hyperparameter optimization algorithms. The experimental results present that our proposed algorithm provides superior results and the design of our proposed algorithm is reasonable. [ABSTRACT FROM AUTHOR]

Disasters have caused significant losses to humans in the past decades. It is essential to learn about the disaster situation so that rescue works can be conducted as soon as possible. Unmanned aerial vehicle (UAV) is a very useful and effective tool to improve the capacity of disaster situational awareness for responders. In the paper, UAV path planning is modelled as the optimization problem, in which fitness functions include travelling distance and risk of UAV, three constraints involve the height of UAV, angle of UAV, and limited UAV slope. An adaptive selection mutation constrained differential evolution algorithm is put forward to solve the problem. In the proposed algorithm, individuals are selected depending on their fitness values and constraint violations. The better the individual is, the higher the chosen probability it has. These selected individuals are used to make mutation, and the algorithm searches around the best individual among the selected individuals. The well-designed mechanism improves the exploitation and maintains the exploration. The experimental results have indicated that the proposed algorithm is competitive compared with the state-of-art algorithms, which makes it more suitable in the disaster scenario. [ABSTRACT FROM AUTHOR]

GloVe representations of words as vector embeddings in continuous spaces are learned from matrix factorization of the words' co-occurrences matrix constructed from large corpora. Due to their high quality as textual features, GloVe embeddings have been extensively utilized for many text mining and natural language processing tasks with considerable success. Further improvements of these word representations can be obtained by also taking into account the valuable information of the semantic properties of the words and the complex relationships between them as provided by semantic lexicons. In this paper we adopt optimization techniques from the domain of machine learning with constrained optimization in order to leverage the relational knowledge between words, and we propose an efficient algorithm that produces word embeddings enhanced by the semantic information. The proposed algorithm outperforms other related approaches that utilize semantic information either during training or as a post-processing step. Our claims are validated by experiments on popular text mining and natural language processing tasks, including word similarities, word analogies, and sentiment analysis, which demonstrate that our proposed model can significantly improve the quality of word vector representations. [ABSTRACT FROM AUTHOR]