A hybrid algorithm for Bayesian network structure learning with application to multi-label learning

We present a novel hybrid algorithm for Bayesian network structure learning, called H2PC. It first reconstructs the skeleton of a Bayesian network and then performs a Bayesian-scoring greedy hill-climbing search to orient the edges. The algorithm is based on divide-and-conquer constraint-based subroutines to learn the local structure around a target variable. We conduct two series of experimental comparisons of H2PC against Max–Min Hill-Climbing (MMHC), which is currently the most powerful state-of-the-art algorithm for Bayesian network structure learning. First, we use eight well-known Bayesian network benchmarks with various data sizes to assess the quality of the learned structure returned by the algorithms. Our extensive experiments show that H2PC outperforms MMHC in terms of goodness of fit to new data and quality of the network structure with respect to the true dependence structure of the data. Second, we investigate H2PC’s ability to solve the multi-label learning problem. We provide theoretical results to characterize and identify graphically the so-called minimal label powersets that appear as irreducible factors in the joint distribution under the faithfulness condition. The multi-label learning problem is then decomposed into a series of multi-class classification problems, where each multi-class variable encodes a label powerset. H2PC is shown to compare favorably to MMHC in terms of global classification accuracy over ten multi-label data sets covering different application domains. Overall, our experiments support the conclusions that local structural learning with H2PC in the form of local neighborhood induction is a theoretically well-motivated and empirically effective learning framework that is well suited to multi-label learning. The source code (in R) of H2PC as well as all data sets used for the empirical tests are publicly available.     

CMOS analogue amplifier circuits optimisation using hybrid backtracking search algorithm with differential evolution

This paper proposes a novel hybrid optimisation algorithm which combines the recently proposed evolutionary algorithm Backtracking Search Algorithm (BSA) with another widely accepted evolutionary algorithm, namely, Differential Evolution (DE). The proposed algorithm called BSA-DE is employed for the optimal designs of two commonly used analogue circuits, namely Complementary Metal Oxide Semiconductor (CMOS) differential amplifier circuit with current mirror load and CMOS two-stage operational amplifier (op-amp) circuit. BSA has a simple structure that is effective, fast and capable of solving multimodal problems. DE is a stochastic, population-based heuristic approach, having the capability to solve global optimisation problems. In this paper, the transistors’ sizes are optimised using the proposed BSA-DE to minimise the areas occupied by the circuits and to improve the performances of the circuits. The simulation results justify the superiority of BSA-DE in global convergence properties and fine tuning ability, and prove it to be a promising candidate for the optimal design of the analogue CMOS amplifier circuits. The simulation results obtained for both the amplifier circuits prove the effectiveness of the proposed BSA-DE-based approach over DE, harmony search (HS), artificial bee colony (ABC) and PSO in terms of convergence speed, design specifications and design parameters of the optimal design of the analogue CMOS amplifier circuits. It is shown that BSA-DE-based design technique for each amplifier circuit yields the least MOS transistor area, and each designed circuit is shown to have the best performance parameters such as gain, power dissipation, etc., as compared with those of other recently reported literature.     

A Markovianity based optimisation algorithm

Several Estimation of Distribution Algorithms (EDAs) based on Markov networks have been recently proposed. The key idea behind these EDAs was to factorise the joint probability distribution of solution variables in terms of cliques in the undirected graph. As such, they made use of the global Markov property of the Markov network in one form or another. This paper presents a Markov Network based EDA that is based on the use of the local Markov property, the Markovianity, and does not directly model the joint distribution. We call it Markovianity based Optimisation Algorithm. The algorithm combines a novel method for extracting the neighbourhood structure from the mutual information between the variables, with a Gibbs sampler method to generate new points. We present an extensive empirical validation of the algorithm on problems with complex interactions, comparing its performance with other EDAs that use higher order interactions. We extend the analysis to other functions with discrete representation, where EDA results are scarce, comparing the algorithm with state of the art EDAs that use marginal product factorisations.     

A hybrid memory-based dragonfly algorithm with differential evolution for engineering application

The dragonfly algorithm (DA) is a swarm-based stochastic algorithm which possesses static and dynamic behavior of swarm and is gaining meaningful popularity due to its low computational cost and fast convergence in solving complex optimization problems. However, it lacks internal memory and is thereby not able to keep track of its best solutions in previous generations. Furthermore, the solution also lacks in diversity and thereby has a propensity of getting trapped in the local optimal solution. In this paper, an iterative-level hybridization of dragonfly algorithm (DA) with differential evolution (DE) is proposed and named as hybrid memory-based dragonfly algorithm with differential evolution (DADE). The reason behind selecting DE is for its computational ability, fast convergence and capability in exploring the solution space through the use of crossover and mutation techniques. Unlike DA, in DADE the best solution in a particular iteration is stored in memory and proceeded with DE which enhances population diversity with improved mutation and accordingly increases the probability of reaching global optima efficiently. The efficiency of the proposed algorithm is measured based on its response to standard set of 74 benchmark functions including 23 standard mathematical benchmark functions, 6 composite benchmark function of CEC2005, 15 benchmark functions of CEC2015 and 30 benchmark function of CEC2017. The DADE algorithm is applied to engineering design problems such as welded beam deign, pressure vessel design, and tension/compression spring design. The algorithm is also applied to the emerging problem of secondary user throughput maximization in an energy-harvesting cognitive radio network. A comparative performance analysis between DADE and other most popular state-of-the-art optimization algorithms is carried out and significance of the results is deliberated. The result demonstrates significant improvement and prominent advantages of DADE compared to conventional DE, PSO and DA in terms of various performance measuring parameters. The results of the DADE algorithm applied on some important engineering design problems are encouraging and validate its appropriateness in the context of solving interesting practical engineering challenges. Lastly, the statistical analysis of the algorithm is also performed and is compared with other powerful optimization algorithms to establish its superiority.

     

A hybrid differential evolution algorithm integrated with an ant system and its application

A new hybrid differential evolution algorithm, in which an ant system is used to select the optimal base vector of mutation operation, named the ant system differential evolution (ASDE), is proposed. In ASDE, each dimension in the feasible solution space is divided into several subspaces evenly, and each subspace is marked with the same initial intensity of pheromone trails. The probability of choosing an individual as the base vector is influenced by the visibility and pheromone quantity of the individual. The trail of the selected base vector’s location subspaces will be reinforced with some pheromones, when the offspring is better than its parent. The experimental results show that the ASDE generally outperforms the other differential evolution algorithms for nine benchmark functions. Furthermore, the ASDE is applied to develop the global kinetic model for SO₂ oxidation on the Cs-Rb-V catalyst, and satisfactory results are obtained.     

A novel two-stage hybrid swarm intelligence optimization algorithm and application

This paper presents a novel two-stage hybrid swarm intelligence optimization algorithm called GA–PSO–ACO algorithm that combines the evolution ideas of the genetic algorithms, particle swarm optimization and ant colony optimization based on the compensation for solving the traveling salesman problem. In the proposed hybrid algorithm, the whole process is divided into two stages. In the first stage, we make use of the randomicity, rapidity and wholeness of the genetic algorithms and particle swarm optimization to obtain a series of sub-optimal solutions (rough searching) to adjust the initial allocation of pheromone in the ACO. In the second stage, we make use of these advantages of the parallel, positive feedback and high accuracy of solution to implement solving of whole problem (detailed searching). To verify the effectiveness and efficiency of the proposed hybrid algorithm, various scale benchmark problems from TSPLIB are tested to demonstrate the potential of the proposed two-stage hybrid swarm intelligence optimization algorithm. The simulation examples demonstrate that the GA–PSO–ACO algorithm can greatly improve the computing efficiency for solving the TSP and outperforms the Tabu Search, genetic algorithms, particle swarm optimization, ant colony optimization, PS–ACO and other methods in solution quality. And the experimental results demonstrate that convergence is faster and better when the scale of TSP increases.     

A memetic particle swarm optimisation algorithm for dynamic multi-modal optimisation problems

Many real-world optimisation problems are both dynamic and multi-modal, which require an optimisation algorithm not only to find as many optima under a specific environment as possible, but also to track their moving trajectory over dynamic environments. To address this requirement, this article investigates a memetic computing approach based on particle swarm optimisation for dynamic multi-modal optimisation problems (DMMOPs). Within the framework of the proposed algorithm, a new speciation method is employed to locate and track multiple peaks and an adaptive local search method is also hybridised to accelerate the exploitation of species generated by the speciation method. In addition, a memory-based re-initialisation scheme is introduced into the proposed algorithm in order to further enhance its performance in dynamic multi-modal environments. Based on the moving peaks benchmark problems, experiments are carried out to investigate the performance of the proposed algorithm in comparison with several state-of-the-art algorithms taken from the literature. The experimental results show the efficiency of the proposed algorithm for DMMOPs.     

A hybrid credibility-based fuzzy multiple objective optimisation to differential pricing and inventory policies with arbitrage consideration

In most markets, price differentiation mechanisms enable manufacturers to offer different prices for their products or services in different customer segments; however, the perfect price discrimination is usually impossible for manufacturers. The importance of accounting for uncertainty in such environments spurs an interest to develop appropriate decision-making tools to deal with uncertain and ill-defined parameters in joint pricing and lot-sizing problems. This paper proposes a hybrid bi-objective credibility-based fuzzy optimisation model including both quantitative and qualitative objectives to cope with these issues. Taking marketing and lot-sizing decisions into account simultaneously, the model aims to maximise the total profit of manufacturer and to improve service aspects of retailing simultaneously to set different prices with arbitrage consideration. After applying appropriate strategies to defuzzify the original model, the resulting non-linear multi-objective crisp model is then solved by a fuzzy goal programming method. An efficient stochastic search procedure using particle swarm optimisation is also proposed to solve the non-linear crisp model.     

A hybrid extreme learning machine model with harris hawks optimisation algorithm: an optimised model for product demand forecasting applications

Accurate and real-time product demand forecasting is the need of the hour in the world of supply chain management. Predicting future product demand from historical sales data is a highly non-linear problem, subject to various external and environmental factors. In this work, we propose an optimised forecasting model - an extreme learning machine (ELM) model coupled with the Harris Hawks optimisation (HHO) algorithm to forecast product demand in an e-commerce company. ELM is preferred over traditional neural networks mainly due to its fast computational speed, which allows efficient demand forecasting in real-time. Our ELM-HHO model performed significantly better than ARIMA models that are commonly used in industries to forecast product demand. The performance of the proposed ELM-HHO model was also compared with traditional ELM, ELM auto-tuned using Bayesian Optimisation (ELM-BO), Gated Recurrent Unit (GRU) based recurrent neural network and Long Short Term Memory (LSTM) recurrent neural network models. Different performance metrics, i.e., Root Mean Squared Error (RMSE), Mean Absolute Percentage Error (MAPE) and Mean Percentage Error (MPE) were used for the comparison of the selected models. Horizon forecasting at 3 days and 7 days ahead was also performed using the proposed approach. The results revealed that the proposed approach is superior to traditional product demand forecasting models in terms of prediction accuracy and it can be applied in real-time to predict future product demand based on the previous week’s sales data. In particular, considering RMSE of forecasting, the proposed ELM-HHO model performed 62.73% better than the statistical ARIMA(7,1,0) model, 40.73% better than the neural network based GRU model, 34.05% better than the neural network based LSTM model, 27.16% better than the traditional non-optimised ELM model with 100 hidden nodes and 11.63% better than the ELM-BO model in forecasting product demand for future 3 months. The novelty of the proposed approach lies in the way the fast computational speed of ELMs has been combined with the accuracy gained by tuning hyperparameters using HHO. An increased number of hyperparameters has been optimised in our methodology compared to available models. The majority of approaches to improve the accuracy of ELM so far have only focused on tuning the weights and the biases of the hidden layer. In our hybrid model, we tune the number of hidden nodes, the number of input time lags and even the type of activation function used in the hidden layer in addition to tuning the weights and the biases. This has resulted in a significant increase in accuracy over previous methods. Our work presents an original way of performing product demand forecasting in real-time in industry with highly accurate results which are much better than pre-existing demand forecasting models.

     

A hybrid clustering algorithm

Clustering attempts to partition a dataset into a meaningful set of mutually exclusive clusters. It is known that sequential clustering algorithms can give optimal partitions when applied to an ordered set of objects. In this technical note, we explore how this approach could be generalized to partition datasets in which there is no natural sequential ordering of the objects. As such, it extends the application of sequential clustering algorithms to all sets of objects.     

A multi-population electromagnetic algorithm for dynamic optimisation problems

This paper is derived from an interest in the development of approaches to tackle dynamic optimisation problems. This is a very challenging research area due to the fact that any approaches utilised should be able to track the changes and simultaneously seek for global optima as the search progresses. In this research work, a multi-population electromagnetic algorithm for dynamic optimisation problems is proposed. An electromagnetic algorithm is a population based meta-heuristic method which imitates the attraction and repulsion of the sample points. In order to track the dynamic changes and to effectively explore the search space, the entire population is divided into several sub-populations (referred as multi-population that acts as diversity mechanisms) where each sub-population takes charge in exploring or exploiting the search space. In addition, further investigation are also conducted on the combination of the electromagnetic algorithm with different diversity mechanisms (i.e. random immigrants, memory mechanism and memory based immigrant schemes) with the aim of identifying the most appropriate diversity mechanism for maintaining the diversity of the population in solving dynamic optimisation problems. The proposed approach has been applied and evaluated against the latest methodologies in reviewed literature of research works with respect to the benchmark problems. This study demonstrates that the electromagnetic algorithm with a multi-population diversity mechanism performs better compared to other population diversity mechanisms investigated in our research and produces some of the best known results when tested on Moving Peak Benchmark (MPB) problems.     

一种基因与蚁群的融合算法研究

蚁群算法具有分布式并行搜索能力，通过信息素的积累和更新收敛于最优路径上，但初期信息素匮乏，收敛较慢。提出一种基因算法与蚁群算法融合的算法，将基因算法加入蚁群算法的每一次迭代中，利用基因算法快速收敛的优点，来加快蚁群系统的收敛速度；且基因算法中的变异机制，有利于提高蚁群算法跳出局部最优的能力。优势互补，实验结果表明该基因蚁群融合算法在寻优能力和收敛速度上都比基因算法和蚁群算法有较大的提高。     

A hybrid approach to artificial bee colony algorithm

In this paper, we put forward a hybrid approach based on the life cycle for the artificial bee colony algorithm to generate dynamical varying population as well as ensure appropriate balance between exploration and exploitation. The bee life-cycle model is firstly constructed, which means that each individual can reproduce or die dynamically throughout the searching process and population size can dynamically vary during execution. With the comprehensive learning, the bees incorporate the information of global best solution into the search equation for exploration, while the Powell’s search enables the bees deeply to exploit around the promising area. Finally, we instantiate a hybrid artificial bee colony (HABC) optimizer based on the proposed model, namely HABC. Comprehensive test experiments based on the well-known CEC 2014 benchmarks have been carried out to compare the performance of HABC against other bio-mimetic algorithms. Our numerical results prove the effectiveness of the proposed hybridization scheme and demonstrate the performance superiority of the proposed algorithm.     

Periodic optimization with applications to solar energy control

It is reasonable to assume in many temperature control applications, that disturbance signals, i.e. solar insolation, ambient temperature, etc. are periodic with 24 hour period and that these periodic disturbances persist over a sufficiently long time interval to allow the system response to settle into a steady-state periodic mode. In this study a linear model is assumed for the system being controlled. The steady-state average of an integral quadratic measure is selected as a performance index. Disturbance signals are assumed to be representable by Fourier series expansions. The periodic optimal control input, which minimizes the given performance index, is then computed in terms of its Fourier coefficient. The optimal control-input coefficients are obtained explicitly in terms of disturbance-input coefficients.     

A new optimization algorithm with application to nonlinear MPC

This paper investigates application of SQP optimization algorithms to nonlinear model predictive control. It considers feasible vs. infeasible path methods, sequential vs. simultaneous methods and reduced vs. full space methods. A new optimization algorithm coined rFOPT which remains feasibile with respect to inequality constraints is introduced. The suitable choices between these various strategies are assessed informally through a small CSTR case study. The case study also considers the effect various discretization methods have on the optimization problem.     

A particle swarm optimisation algorithm with interactive swarms for tracking multiple targets

We propose a novel particle swarm optimisation algorithm that uses a set of interactive swarms to track multiple pedestrians in a crowd. The proposed method improves the standard particle swarm optimisation algorithm with a dynamic social interaction model that enhances the interaction among swarms. In addition, we integrate constraints provided by temporal continuity and strength of person detections in the framework. This allows particle swarm optimisation to be able to track multiple moving targets in a complex scene. Experimental results demonstrate that the proposed method robustly tracks multiple targets despite the complex interactions among targets that lead to several occlusions.     

A hybrid grasshopper optimization algorithm with bat algorithm for global optimization

Multimedia Tools and Applications - This paper introduces a hybrid grasshopper optimization algorithm with bat algorithm (BGOA) for global optimization. In the BGOA, the Levy flight with variable...     

A hybrid genetic algorithm with the Baldwin effect

Here we present a new hybrid genetic algorithm (HGA) with the Baldwin effect. In the HGA, a local search is employed to change the fitness of individuals but the acquired improvements do not change the individual itself. This local search step exploits the Baldwin effect. Some numerical applications show that this algorithm can yield the global optimum more efficiently than commonly used HGAs. A theorem is presented that guarantees the convergence in probability of the new HGA.     

A hybrid multi-objective evolutionary algorithm with feedback mechanism

Exploration and exploitation are two cornerstones for multi-objective evolutionary algorithms (MOEAs). To balance exploration and exploitation, we propose an efficient hybrid MOEA (i.e., MOHGD) by integrating multiple techniques and feedback mechanism. Multiple techniques include harmony search, genetic operator and differential evolution, which can improve the search diversity. Whereas hybrid selection mechanism contributes to the search efficiency by integrating the advantages of the static and adaptive selection scheme. Therefore, multiple techniques based on the hybrid selection strategy can effectively enhance the exploration ability of the MOHGD. Besides, we propose a feedback strategy to transfer some non-dominated solutions from the external archive to the parent population. This feedback strategy can strengthen convergence toward Pareto optimal solutions and improve the exploitation ability of the MOHGD. The proposed MOHGD has been evaluated on benchmarks against other state of the art MOEAs in terms of convergence, spread, coverage, and convergence speed. Computational results show that the proposed MOHGD is competitive or superior to other MOEAs considered in this paper.