Stochastic Search Strategy Research Articles

An improved artificial bee colony algorithm based on elite group guidance and combined breadth-depth search strategy

Inspired by the natural phenomenon that honey bees follow the elite group in the foraging process, we propose a novel artificial bee colony algorithm named ECABC based on elite group guidance and the combined breadth-depth search strategy in this paper. Firstly, by simulating the behavior that bees follow the elite group, a novel neighborhood search equation is proposed. According to the equation, with the center of the elite group as the starting point of the search, under the guidance of the global optimum, neighborhood search is performed. Secondly, a combined search strategy is designed as follows: the stochastic breadth-first search strategy for employed bees and the stochastic depth-first search strategy for onlooker bees. Thirdly, the random selection method of elite bees is adopted to replace the probability selection method of onlooker bees. Besides, the influencing parameters of the optimization results are studied and the optimum parameters allowing the best comprehensive performance are obtained. In addition, the proposed algorithm is experimentally verified with 22 benchmark functions and then compared with other improved artificial bee colony algorithms. The comparison results show that the ECABC can effectively improve the convergence speed, convergence precision, and robustness.

Role of the vibrational contribution in Coulomb explosion of dicationic neon gas clusters: a parallel tempering based study.

The problem of Coulomb explosion in dicationic neon gas clusters has been investigated with special emphasis on the role of the vibrational contribution. The problem has been handled by describing the dicationic neon gas system with an adequate potential energy surface comprising dispersive interaction, Coulombic and polarizability containing terms. This potential energy surface, if explored for various sizes of the clusters, shows Coulombic explosion features below a certain threshold size. However this classical treatment fails to account for the correct threshold predicted from other studies including experiments. This signifies that quantum effects play an important role. With the incorporation of the vibrational contribution as the quantum effect, it is seen that reduction in the threshold value indeed occurs and the amount of decrease significantly varies with temperature. The whole study has been done using the stochastic search strategy or parallel tempering to explore the potential energy surface of the system. The stochastic strategy guarantees the achievement of a low energy solution as it is not stuck in local energy basins.

Statistical signatures of a targeted search by bacteria

Chemoattractant gradients are rarely well-controlled in nature and recent attention has turned to bacterial chemotaxis toward typical bacterial food sources such as food patches or even bacterial prey. In environments with localized food sources reminiscent of a bacterium’s natural habitat, striking phenomena—such as the volcano effect or banding—have been predicted or expected to emerge from chemotactic models. However, in practice, from limited bacterial trajectory data it is difficult to distinguish targeted searches from an untargeted search strategy for food sources. Here we use a theoretical model to identify statistical signatures of a targeted search toward point food sources, such as prey. Our model is constructed on the basis that bacteria use temporal comparisons to bias their random walk, exhibit finite memory and are subject to random (Brownian) motion as well as signaling noise. The advantage with using a stochastic model-based approach is that a stochastic model may be parametrized from individual stochastic bacterial trajectories but may then be used to generate a very large number of simulated trajectories to explore average behaviors obtained from stochastic search strategies. For example, our model predicts that a bacterium’s diffusion coefficient increases as it approaches the point source and that, in the presence of multiple sources, bacteria may take substantially longer to locate their first source giving the impression of an untargeted search strategy.

Constructing a bunch of paths for conformational changes in size specific noble gas cluster: a study using a stochastic procedure

The problem of constructing reaction paths for conformational changes in noble gas clusters, which are formed by weak dispersion forces has been discussed. The potential energy surface describing these systems are extremely rugged and contains a large number of minima and saddle points within a narrow energy range. So it is not beyond expectation that there must exist a number of paths describing conformational changes in these systems in the given energy range and these bunch of paths provide a better description of possible conformational changes as opposed to the conventional single minimum energy path. To study this issue we have taken up the $$Ar_{18}$$ cluster as a typical example and the entire formulation is done with the use of the stochastic search strategy of simulated annealing. The quantitative results clearly show the existence of a large number of minima and transition states and also a multitude of probable transforming paths between stable geometries.

Decision Support in Intelligent Maintenance-planning Systems Based on Contextual Multi-armed Bandit Algorithm

In this paper we focus on two essential problems of maintenance decision support systems, namely, 1) detection of potential dangerous situation, and 2) classification of this situation in order to recommend an appropriate repair action. The former task is usually solved with the known statistical process control techniques. The latter problem can be reduced to the contextual multi- armed bandit problem. We propose a novel algorithm with Bayesian classification of abnormal situation and the softmax rule to explore the decision space. The dangerous situations are detected with the Shewhart control charts for the distances between the current and the normal situations. It is experimentally shown, that our algorithm is more accurate than the known contextual multi-armed methods with stochastic search strategies.

Exploring Dynamic Environments Using Stochastic Search Strategies

In this paper, we conduct a literature review of laws of motion based on stochastic search strategies which are mainly focused on exploring highly dynamic environments. In this regard, stochastic search strategies represent an interesting alternative to cope with uncertainty and reduced perceptual capabilities. This study aims to present an introductory overview of research in terms of directional rules and searching methods mainly based on bio-inspired approaches. This study critically examines the role of animal searching behavior applied to random walk models using stochastic rules and kinesis or taxis. The aim of this study is to examine existing techniques and to select relevant work on random walks and analyze their actual contributions. In this regard, we cover a wide range of displacement events with an orientation mechanism given by a reactive behavior or a source-seeking behavior. Finally, we conclude with a discussion concerning the usefulness of using optimal foraging strategies as a reliable methodology.

A stochastically evolving non-local search and solutions to inverse problems with sparse data

Building on a martingale approach to global optimization, a powerful stochastic search scheme for the global optimum of cost functions is proposed using change of measures on the states that evolve as diffusion processes and splitting of the state-space along the lines of a Bayesian game. To begin with, the efficacy of the optimizer, when contrasted with one of the most efficient existing schemes, is assessed against a family of Np-hard benchmark problems. Then, using both simulated and experimental data, potentialities of the new proposal are further explored in the context of an inverse problem of significance in photoacoustic imaging, wherein the superior reconstruction features of a global search vis-à-vis the commonly adopted local or quasi-local schemes are brought into relief.

Folding of small knotted proteins: Insights from a mean field coarse-grained model.

A small but relevant number of proteins whose native structure is known features nontrivial topology, i.e., they are knotted. Understanding the process of folding from a swollen unknotted state to the biologically relevant native conformation is, for these proteins, particularly difficult, due to their rate-limiting topological entanglement. To shed some light into this conundrum, we introduced a structure-based coarse-grained model of the protein, where the information about the folded conformation is encoded in bonded angular interactions only, which do not favor the formation of native contacts. A stochastic search scheme in parameter space is employed to identify a set of interactions that maximizes the probability to attain the knotted state. The optimal knotting pathways of the two smallest knotted proteins, obtained through this approach, are consistent with the results derived by means of coarse-grained as well as full atomistic simulations.

An optimisation approach for culturing shear-sensitive dinoflagellate microalgae in bench-scale bubble column photobioreactors

The dinoflagellate Karlodinium veneficum was grown in bubble column photobioreactors and a genetic algorithm-based stochastic search strategy used to find optimal values for the culture parameters gas flow rate, culture height, and nozzle sparger diameter. Cell production, concentration of reactive oxygen species (ROS), membrane fluidity and photosynthetic efficiency were studied throughout the culture period. Gas-flow rates below 0.26Lmin−1, culture heights over 1.25m and a nozzle diameter of 1.5mm were found to provide the optimal conditions for cell growth, with an increase of 60% in cell production with respect to the control culture. Non-optimal conditions produced a sufficiently high shear stress to negatively affect cell growth and even produce cell death. Cell physiology was also severely affected in stressed cultures. The production of ROS increased by up to 200%, whereas cell membrane fluidity decreased by 60% relative to control cultures. Photosynthetic efficiency decreased concomitantly with membrane fluidity.

An optimal culture medium for growing Karlodinium veneficum: Progress towards a microalgal dinoflagellate-based bioprocess

In addressing a potential bioprocess for bioactives production from microalgae, the choice of a proper culture medium is a key factor. In the present work, an optimal culture medium for producing the marine dinoflagellate microalga Karlodinium veneficum was formulated using a stochastic search strategy based on genetic algorithm. The optimized formulation had 25 components including the macronutrients, trace elements and vitamins. With this medium, the final attainable cell concentration during batch growth was enhanced by 120% relative to the control culture grown in the L1 medium. The final titer of the bioactive karlotoxins measured using the hemolytic activity of the culture filtrate was enhanced by 190% relative to the L1 control. These results are a promising basis for further development of photobioreactor culture strategies for microalgal dinoflagellates.

On Decentralized Estimation With Active Queries

We consider the problem of decentralized 20 questions with noise for multiple players/agents under the minimum entropy criterion in the setting of stochastic search over a parameter space, with application to target localization. We propose decentralized extensions of the active query-based stochastic search strategy that combines elements from the 20 questions approach and social learning. We prove convergence to correct consensus on the value of the parameter. This framework provides a flexible and tractable mathematical model for decentralized parameter estimation systems based on active querying. We illustrate the effectiveness and robustness of the proposed decentralized collaborative 20 questions algorithm for random network topologies with information sharing.

Stochastic optimal foraging: tuning intensive and extensive dynamics in random searches.

Recent theoretical developments had laid down the proper mathematical means to understand how the structural complexity of search patterns may improve foraging efficiency. Under information-deprived scenarios and specific landscape configurations, Lévy walks and flights are known to lead to high search efficiencies. Based on a one-dimensional comparative analysis we show a mechanism by which, at random, a searcher can optimize the encounter with close and distant targets. The mechanism consists of combining an optimal diffusivity (optimally enhanced diffusion) with a minimal diffusion constant. In such a way the search dynamics adequately balances the tension between finding close and distant targets, while, at the same time, shifts the optimal balance towards relatively larger close-to-distant target encounter ratios. We find that introducing a multiscale set of reorientations ensures both a thorough local space exploration without oversampling and a fast spreading dynamics at the large scale. Lévy reorientation patterns account for these properties but other reorientation strategies providing similar statistical signatures can mimic or achieve comparable efficiencies. Hence, the present work unveils general mechanisms underlying efficient random search, beyond the Lévy model. Our results suggest that animals could tune key statistical movement properties (e.g. enhanced diffusivity, minimal diffusion constant) to cope with the very general problem of balancing out intensive and extensive random searching. We believe that theoretical developments to mechanistically understand stochastic search strategies, such as the one here proposed, are crucial to develop an empirically verifiable and comprehensive animal foraging theory.

Differential evolution algorithm for nonlinear inversion of high-frequency Rayleigh wave dispersion curves

Abstract In recent years, Rayleigh waves are gaining popularity to obtain near-surface shear (S)-wave velocity profiles. However, inversion of Rayleigh wave dispersion curves is challenging for most local-search methods due to its high nonlinearity and to its multimodality. In this study, we proposed and tested a new Rayleigh wave dispersion curve inversion scheme based on differential evolution (DE) algorithm. DE is a novel stochastic search approach that possesses several attractive advantages: (1) Capable of handling non-differentiable, non-linear and multimodal objective functions because of its stochastic search strategy; (2) Parallelizability to cope with computation intensive objective functions without being time consuming by using a vector population where the stochastic perturbation of the population vectors can be done independently; (3) Ease of use, i.e. few control variables to steer the minimization/maximization by DE's self-organizing scheme; and (4) Good convergence properties. The proposed inverse procedure was applied to nonlinear inversion of fundamental-mode Rayleigh wave dispersion curves for near-surface S-wave velocity profiles. To evaluate calculation efficiency and stability of DE, we firstly inverted four noise-free and four noisy synthetic data sets. Secondly, we investigated effects of the number of layers on DE algorithm and made an uncertainty appraisal analysis by DE algorithm. Thirdly, we made a comparative analysis with genetic algorithms (GA) by a synthetic data set to further investigate the performance of the proposed inverse procedure. Finally, we inverted a real-world example from a waste disposal site in NE Italy to examine the applicability of DE on Rayleigh wave dispersion curves. Furthermore, we compared the performance of the proposed approach to that of GA to further evaluate scores of the inverse procedure described here. Results from both synthetic and actual field data demonstrate that differential evolution algorithm applied to nonlinear inversion of high-frequency surface wave data should be considered good not only in terms of the accuracy but also in terms of the convergence speed.

A particle swarm optimization using local stochastic search and enhancing diversity for continuous optimization

The particle swarm optimizer (PSO) is a swarm intelligence based on heuristic optimization technique that can be applied to a wide range of problems. After analyzing the dynamics of traditional PSO, this paper presents a new PSO variant on the basis of local stochastic search strategy (LSSPSO) for performance improvement. This is encouraged by a social phenomenon that everyone wants to first exceed the nearest superior and then all superior. Specifically, LSSPSO employs a local stochastic search to adjust inertia weight in terms of keeping a balance between the diversity and the convergence speed, aiming to improve the performance of traditional PSO. Experiments conducted on unimodal and multimodal test functions demonstrate the effectiveness of LSSPSO in solving multiple benchmark problems as compared to several other PSO variants.

SAWSA-LPR: Astochastic search strategy for estimation of maximum likelihood DNA phylogenetic trees

In the spirit of the “grand challenge”, this paper covers the development of novel concepts for inference of large phylogenies based on the maximum likelihood method, which has proved to be the most accurate model for inference of huge and complex phylogenetic trees. Here, a novel method called Leaf Pruning and Re-grafting (LPR) has being presented, which is a variant of standard Sub-tree Pruning and Re-grafting (SPR) technique. LPR is a systematic approach where only unique topologies are generated at each step. Various stochastic search strategies for estimation of the maximum likelihood (ML) tree have also being proposed. Here, simulated annealing has been combined with steepest accent method to improve the quality of the final tree obtained. All the current simulated annealing approaches are used with simple hill climbing method to avoid the large number of repeated topologies that are normally generated by SPR. This easily leads to local maxima. However in the present study steepest accent with simulated annealing by way of LPR (SAWSA-LPR) has being used; the chances of returning local maxima has being significantly reduced. A straightforward and efficient parallel version of simulated annealing with steepest accent to accelerate the process of DNA phylogenetic tree inference has also being presented. It was observed that the implementation of the algorithm based on random DNA sequences gave better results as compared to other tree construction methods.

Structural and spectroscopic studies of carbon dioxide clusters: a combined genetic algorithm and DFT based study

In this manuscript we use a two-tier approach of combining stochastic search strategies and DFT calculations to arrive at extremely good quality structures for neutral (CO2) n clusters with n = 2–13. The idea of using stochastic search strategy initially is to find out extreme quickly, that is with low computational cost, more than one reasonably good structures by exploring an empirical potential energy surface for CO2 clusters. These pre-optimized structures [using genetic algorithm (GA) as stochastic optimizer] if used as inputs for subsequent DFT calculations using dispersion-corrected ωB97XD and meta hybrid M06-2X functionals give global minimum structures for these systems. This two-step strategy in our opinion is a more potent route to obtain the global minimum structures with certainty rather than a dedicated quantum chemical calculation carried from the outset. We further analyse in detail the structural, vibrational spectra and interaction energies of the clusters studied. We compare our results with experimental and theoretical results available in literature. The correspondence with experimental results is nearly 99 %, that is the results show marginal deviation from reported experimental vibrational frequencies and hence we conclude that GA in conjunction with dispersion-corrected DFT is a viable method for the study of clusters where the interaction is non-covalent in nature. Since B2PLYP-D is also an efficient functional in the study of these systems, we also report interaction energies of the cluster systems examined using this functional.

A discrete artificial bee colony for multiple Knapsack problem

Multiple Knapsack problem (MKP) is a most popular multiple subset selection problem that belongs to the class of NP–Complete problems. The aim is to assign optimal subsets among all original items to some knapsacks, such that the overall profit of all selected items be maximised, while the total weight of all assigned items to any knapsack does not exceed the allowable capacity of it. Artificial bee colony (ABC) algorithm is a new meta–heuristic with a stochastic search strategy. In ABC, the neighbourhood area of any best–found solution is searched by the employed bees to achieve better solutions. This paper presents a discrete ABC algorithm for the MKP. In this approach, a hybrid probabilistic mutation scheme is performed for searching the neighbourhood of food sources. The proposed algorithm can guide the search space quickly and improve the local search ability. Experimental results demonstrate that the presented approach has improved the quality and convergence speed than other evolutionary algorithms.

McGenus: a Monte Carlo algorithm to predict RNA secondary structures with pseudoknots

We present McGenus, an algorithm to predict RNA secondary structures with pseudoknots. The method is based on a classification of RNA structures according to their topological genus. McGenus can treat sequences of up to 1000 bases and performs an advanced stochastic search of their minimum free energy structure allowing for non-trivial pseudoknot topologies. Specifically, McGenus uses a Monte Carlo algorithm with replica exchange for minimizing a general scoring function which includes not only free energy contributions for pair stacking, loop penalties, etc. but also a phenomenological penalty for the genus of the pairing graph. The good performance of the stochastic search strategy was successfully validated against TT2NE which uses the same free energy parametrization and performs exhaustive or partially exhaustive structure search, albeit for much shorter sequences (up to 200 bases). Next, the method was applied to other RNA sets, including an extensive tmRNA database, yielding results that are competitive with existing algorithms. Finally, it is shown that McGenus highlights possible limitations in the free energy scoring function. The algorithm is available as a web server at http://ipht.cea.fr/rna/mcgenus.php.

Stochastic search strategies in 2D using agents with limited perception

Animals use stochastic search strategies to find a potential nest or to localize the food source. Animals do not rely much on the environmental cues or their sensory inputs. Since their cognitive capabilities are rather limited, search strategies are usually simplistic and stochastic. And yet such simple search strategies result in high efficiency in finding potential targets in limited time intervals. Similar problems with perception of the environment and limited computational power and sensory inputs are observed on the mobile robots used in search scenarios. In spite of their limitations, such robotic systems usually perform deterministic search strategies which heavily depend on environmental cues and sensory input. The research question we are addressing in this paper is which stochastic search strategy has the best performance in terms of area coverage and can stochastic search strategies outperform deterministic search strategies.

Genetic algorithm-based medium optimization for a toxic dinoflagellate microalga

A genetic algorithm stochastic search strategy was used to optimize the culture medium for producing the toxic marine dinoflagellate microalga Protoceratium reticulatum. The optimized medium contained 26 different components (macronutrients, trace elements, vitamins). The use of this medium allowed a 60% enhancement in the final cell concentration relative to the control culture based on L1 medium. The final titer of yessotoxins was improved by 40% relative to the control medium.