Dynamic Search Space Research Articles

Collaborative optimization decision making of cement grinding process operational indicators considering dual dynamic problem

The cement grinding process requires monitoring of unit power consumption and specific surface area indicators to improve production efficiency and product quality. The strong coupling between operational variables necessitates careful adjustment of these variables to prevent production from becoming instability. Furthermore, the continuous and time-dependent nature of the process makes this adjustment particularly challenging. To address this dual dynamic problem, we have developed a multi-objective optimization model, with global optimization as the desired outcome, to optimize unit power consumption and specific surface area. Our optimization algorithm, termed “Optimization Algorithm - Dynamic Search Space and Rolling Time Domain” (OA-DSTR), considers the problem of dynamic production process. First, the algorithm uses a dynamic search space strategy, allowing for changes in the constraint range of the operational variables and introducing a fluctuation coefficient (FC) to measure solution rationality. Second, to deal with the time-dependent problem, we have added a rolling time domain strategy, enabling real-time monitoring of the cement grinding process. The experimental results show that OA-DSTR not only ensures global optimization, but also realizes the tracking of dynamic conditions, improving the stability of the cement grinding process and achieving a lower FC value.

Dynamic Random Walk and Dynamic Opposition Learning for Improving Aquila Optimizer: Solving Constrained Engineering Design Problems.

One of the most important tasks in handling real-world global optimization problems is to achieve a balance between exploration and exploitation in any nature-inspired optimization method. As a result, the search agents of an algorithm constantly strive to investigate the unexplored regions of a search space. Aquila Optimizer (AO) is a recent addition to the field of metaheuristics that finds the solution to an optimization problem using the hunting behavior of Aquila. However, in some cases, AO skips the true solutions and is trapped at sub-optimal solutions. These problems lead to premature convergence (stagnation), which is harmful in determining the global optima. Therefore, to solve the above-mentioned problem, the present study aims to establish comparatively better synergy between exploration and exploitation and to escape from local stagnation in AO. In this direction, firstly, the exploration ability of AO is improved by integrating Dynamic Random Walk (DRW), and, secondly, the balance between exploration and exploitation is maintained through Dynamic Oppositional Learning (DOL). Due to its dynamic search space and low complexity, the DOL-inspired DRW technique is more computationally efficient and has higher exploration potential for convergence to the best optimum. This allows the algorithm to be improved even further and prevents premature convergence. The proposed algorithm is named DAO. A well-known set of CEC2017 and CEC2019 benchmark functions as well as three engineering problems are used for the performance evaluation. The superior ability of the proposed DAO is demonstrated by the examination of the numerical data produced and its comparison with existing metaheuristic algorithms.

Correction to: Design Optimization of Water Distribution Networks with Dynamic Search Space Reduction GA

Correction to: Design Optimization of Water Distribution Networks with Dynamic Search Space Reduction GA

Design Optimization of Water Distribution Networks with Dynamic Search Space Reduction GA

Design Optimization of Water Distribution Networks with Dynamic Search Space Reduction GA

Enhanced Ant Colony Optimization for Vehicular Ad Hoc Networks Using Fittest Node Clustering

Vehicular ad hoc networks (VANETs) are a rapidly evolving field at the intersection of intelligent transportation systems, emphasizing the need for a stable and scalable VANET topology to accommodate growing vehicular densities. The intricate challenge of route selection calls for advanced clustering protocols to bolster road safety and message routing. This research introduces a novel approach to intelligent clustering routing protocols, leveraging heuristic-based solutions built upon an enhanced ant colony optimizer (ACO) framework. The study unfolds in two stages: the creation of a dynamic search space model and the election of cluster heads (CHs). The innovative dynamic aware transmission range parallel Euclidean distance (DA-TRPED) technique establishes a dynamic search space using the parallel Euclidean distance (PED) concept. This approach evaluates vehicular nodes by estimating PED values, reducing the search process’s complexity. Subsequently, an intelligent cluster head is selected by enhancing the dynamic evaporation factor (DEF) within the ACO technique. The experimental validation of the DA-TRPED technique takes place in NS2 simulations, demonstrating superior performance compared to conventional ACO. This enhancement is evident in metrics such as packet delivery, packet drop, throughput, end-to-end delay, and the lifetime analysis of clustered nodes. The proposed approach holds promise for optimizing VANETs, enhancing their stability and scalability while promoting road safety and efficient message routing.

Boosting autonomous process design and intensification with formalized domain knowledge

Conceptual process design deals with searching for optimal process flowsheets in a large design space. Effective approaches benefit from sensible search space restrictions, commonly carried out by a knowledgeable expert in the form of a superstructure optimization or heuristic rules. To achieve the goal of autonomous process design, knowledge has to be formalized in a machine-readable format. This contribution aims to incorporate an ontological representation of fundamental process knowledge to empower general-purpose design procedures while respecting problem-specific variability. Specifically, the presented framework leverages an ontology to express declarative knowledge (what-is) of processes, phenomena and design tasks to set up the search space and boost a hierarchical reinforcement learning agent which learns the required procedural knowledge (how-to) in order to find an optimal solution. The work is applied in a case study of an intensified steam methane reforming process. Results show that the automated treatment of domain knowledge allows for dynamic search space reduction and achieves better computational efficiency and solution quality, highlighting its potential in autonomous process design approaches.

Enhanced Anomaly Detection System for IoT Based on Improved Dynamic SBPSO.

The Internet of Things (IoT) supports human endeavors by creating smart environments. Although the IoT has enabled many human comforts and enhanced business opportunities, it has also opened the door to intruders or attackers who can exploit the technology, either through attacks or by eluding it. Hence, security and privacy are the key concerns for IoT networks. To date, numerous intrusion detection systems (IDS) have been designed for IoT networks, using various optimization techniques. However, with the increase in data dimensionality, the search space has expanded dramatically, thereby posing significant challenges to optimization methods, including particle swarm optimization (PSO). In light of these challenges, this paper proposes a method called improved dynamic sticky binary particle swarm optimization (IDSBPSO) for feature selection, introducing a dynamic search space reduction strategy and a number of dynamic parameters to enhance the searchability of sticky binary particle swarm optimization (SBPSO). Through this approach, an IDS was designed to detect malicious data traffic in IoT networks. The proposed model was evaluated using two IoT network datasets: IoTID20 and UNSW-NB15. It was observed that in most cases, IDSBPSO obtained either higher or similar accuracy even with less number of features. Moreover, IDSBPSO substantially reduced computational cost and prediction time, compared with conventional PSO-based feature selection methods.

One-shot Graph Neural Architecture Search with Dynamic Search Space

Relying on the diverse graph convolution operations that have emerged in recent years, graph neural networks (GNNs) are shown to be powerful to deal with high-dimensional non-Euclidean domains, such as social networks or citation networks. Despite the tremendous human efforts been taken to explore new graph convolution operations, there are a few attempts to automatically search operations in GNNs. The search space of GNNs is significantly larger than that of CNNs, because of diverse components in the message-passing of GNNs. This, therefore, prevents the straightforward application of classical NAS methods for GNNs. In this work, we propose a novel dynamic one-shot search space for multi-branch neural architectures of GNNs. The dynamic search space maintains a subset of the large search space along with a set of importance weights for operation candidates in the subset as the architecture parameters. After each iteration, the subset is pruned by removing candidates with low importance weights and is expanded with new operations. The dynamic subsets of operation candidates are not uniform but is individual for each edge in the computation graph of the neural architecture, which can ensure the diversity of operations in the final architecture is as competitive as direct search in the large search space. Our experiments of semi-supervised and supervised node classification on citation networks, including Cora, Citeseer, and Pubmed, demonstrate that our method outperforms the current state-of-the-art manually designed architectures and reaches competitive performance to existing GNN NAS approaches with up to 10 times of speedup.

An adaptive regeneration framework based on search space adjustment for differential evolution

Differential evolution (DE) is a well-known evolutionary algorithm with simple operation and excellent performance, which has been applied to solve various optimization problems. To alleviate the problem of premature convergence or population stagnation faced by DE algorithm, this paper proposes an adaptive regeneration framework based on search space adjustment (ARSA), which can be easily embedded into various DE variants. When one individual cannot get improved for several generations, the ARSA framework will be triggered to randomly generate a substitute individual from a dynamic search space which is determined by a given adjustment mechanism controlled by two different levels of parameters. The trigger condition for each individual is adaptively controlled by its status in the current population. The space adjustment mechanism contains two strategies, one focuses on global exploration while the other on local exploitation. Moreover, the ARSA framework does not add any parameters that need to be pre-set, and all the included parameters are adaptive. To verify the availability of ARSA framework for solving complex optimization problems, thirty functions with different dimensions from IEEE CEC 2017 test platform and three real-life problems are employed for comparative experiments. The experimental results indicate that our ARSA framework notably improves the performance of two basic DE algorithms and six state-of-the-art DE variants.

An enhanced differential evolution algorithm with a new oppositional-mutual learning strategy

Global optimization has been a hot research topic in various engineering applications, where differential evolution (DE) is one of the most popular approaches. Actually, it is inevitable for DE to trap into local optima when dealing with complex optimization problems. Dynamic opposite learning (DOL), which is a new variant of opposition-based learning (OBL), has the potential for enhancing DE, due to its strong exploration capability contributed by the asymmetric and dynamic search space. To balance exploration and exploitation, an adjustable weight parameter of the search space is adopted, yet adjusting the value needs a lot of tests and expert experience. Instead of relying on the additional weight parameter, a mutual learning (ML) strategy, which leads individuals to learn from each other deterministically, is combined with DOL for raising exploitation. The trade-off between exploration and exploitation is guaranteed by randomly switching DOL and ML in the population initialization process and the generation jumping process. A hybrid strategy, named oppositional-mutual learning (OML), is thereby generated, and it is applied for the performance improvement of DE. Benchmarks from CEC 2014, including unimodal, multi-modal, hybrid and composition functions, were adopted to evaluate the performance of the oppositional-mutual learning DE (OMLDE). Numerical results with the comparisons to the state-of-the-art counterparts show that OMLDE has significant advantages of converging to the global optimum on most functions, which also validates the superiority of the novel OML strategy.

An improved antlion optimizer with dynamic random walk and dynamic opposite learning

Antlion optimization algorithm (ALO) has been one of the most popular meta-heuristic algorithms to solve engineering optimization problems with a fast convergence speed and steady performance. However, the long runtime and the weak capability of local optima avoidance prevent ALO from being applied in real-time and complex engineering problems. This paper proposes a dynamic random walk and dynamic opposite learning improved ALO (DALO), where a new random walk method named dynamic random walk (DRW) is presented, and a dynamic opposite learning strategy (DOL) is embedded in ALO. The DOL-inspired DRW method has a better exploration capability for converging to the best optimum due to the dynamic changing search space and is more efficient in computation due to the low complexity. The DOL strategy contributes to two steps of the ALO algorithm, i.e. DOL population initialization and DOL generation jumping, for further enhancing the algorithm to avoid premature convergence. The improvements of DRW and DOL were verified by testing benchmark problems from ALO paper, CEC 2014 and CEC 2017, and an optimal weight factor of DRW was selected by conducting a sensitivity analysis. The results, confirmed by Wilcoxon rank-sum test and Friedman test, show that DRW helps ALO greatly reduce the runtime with an excellent performance in finding the best optimum, and the DALO algorithm with the further enhancement enable by DOL is significantly superior to ALO. In addition, the results of the test on three representative engineering problems show that the ability of DALO to solve real-world problems is enhanced.

Application of Dynamically Search Space Squeezed Modified Firefly Algorithm to a Novel Short Term Economic Dispatch of Multi-Generation Systems

The absence of the global best component in the update equation of the conventional firefly algorithm degrades its exploration properties. This research proposes multi-update position criteria to enhance the exploration properties of the conventional firefly technique while including the effect of the global best solution on the movement of the fireflies in the search space of the objective function. Moreover, the dynamic search space squeezing is applied to constrict the movement of the fireflies within the certain limits to avoid their oscillatory movement as the solution approaches towards the global best by determining the optimal trajectory for each firefly. The robustness of the suggested firefly algorithm is tested on a hybrid energy system consisting of thermal, hydroelectric, and Photovoltaic (PV) energy source. The intermittent nature of the PV energy source is explained using fractional integral polynomial model and Auto Regressive Integrated Moving Average (ARIMA) model. The main dispatch problem is successfully computed using both the modified firefly and the simple firefly algorithm by determining the optimal power share of each energy source for different scheduling intervals. The suggested operational strategy reduces the overall generation cost of the system while preserving the various system constraints. Due to the stochastic nature of the meta-heuristic techniques, the two suggested algorithms are compared statistically for different test cases using the independent t-test results. The statistical comparison suggests that the performance of the modified firefly is superior to its conventional counterpart as the evaluation parameters of the modified firefly converge to relatively lower value as compared to the parameters of the simple firefly algorithm.

An adaptive dimension level adjustment framework for differential evolution

Differential evolution (DE) has been recognized as one of the most popular evolutionary algorithms. There are numerous DE variants adopting multi-operators based cooperation strategy to improve their performance, but almost all of the adopted cooperation strategies are essentially implemented at the individual level or population level, and the implementation at the dimension level are scarce. In this paper, an adaptive dimension level adjustment (ADLA) framework is designed to relieve the premature convergence or stagnation problem faced by DE algorithm, which can be easily combined with diverse DE variants. When the current optimal individual cannot get improved for a given uninterrupted iterations, ADLA framework will be triggered to select some individuals at random according to specific rule and reinitialize portion of their dimensions from a dynamic search space that adjusted by a population level macroparameter and one individual level microparameter. Moreover, ADLA framework contains two reinitialization operators with different search characteristics, and the coordination between them is executed at the dimension level, which has potential advantages in balancing the global exploration ability and local exploitation ability. Extensive comparison experiments are carried out based on IEEE CEC 2014 test platform, two basic DE algorithms and six outstanding DE variants. The experimental results demonstrate that ADLA framework can memorably enhance the performance of every DE algorithm used for comparison.

Ship Motion Attitude Prediction Based on an Adaptive Dynamic Particle Swarm Optimization Algorithm and Bidirectional LSTM Neural Network

A new neural network prediction model is proposed for predicting ship motion attitude with high accuracy. This prediction model is based on an adaptive dynamic particle swarm optimization algorithm (ADPSO) and bidirectional long short-term memory (BiLSTM) neural network, which is to optimize the hyperparameters of BiLSTM neural network by the proposed ADPSO algorithm. The ADPSO algorithm introduces dynamic search space strategy into the classical particle swarm optimization algorithm and adjusts the learning factor adaptively to balance the global and local search ability, so as to improve the optimization performance and improve its optimization effect in BiLSTM parameter optimization process. The results show that the model can obtain higher prediction accuracy and faster convergence speed, and has better prediction performance in the prediction of ship motion attitude.

Statistical Performances Evaluation of APSO and Improved APSO for Short Term Hydrothermal Scheduling Problem

The Accelerated Particle Swarm Optimization (APSO) algorithm is an efficient and the easiest to implement variant of the famous Particle Swarm Optimization (PSO) algorithm. PSO and its variant APSO have been implemented on the famous Short-Term Hydrothermal Scheduling (STHTS) problem in recent research, and they have shown promising results. The APSO algorithm can be further modified to enhance its optimizing capability by deploying dynamic search space squeezing. This paper presents the implementation of the improved APSO algorithm that is based on dynamic search space squeezing, on the short-term hydrothermal scheduling problem. To give a quantitative comparison, a true statistical comparison based on comparing means is also presented to draw conclusions.

An improved evolutionary approach-based hybrid algorithm for Bayesian network structure learning in dynamic constrained search space

Learning Bayesian network (BN) structures from data is a NP-hard problem due to the vastness of the solution space. To address this issue, hybrid approaches that integrate the constraint-based (CB) method and the score-and-search (SS) method have been developed in the literature, but when the constrained search space is fixed and inaccurate, it is very likely to lose the optimal solution, leading to low learning accuracy. Besides, due to the randomness and uncertainty of the search, it is difficult to preserve the superiority of the structures, resulting in low learning efficiency. Therefore, we propose a novel hybrid algorithm based on an improved evolutionary approach to explore BN structure with highest matching degree of data set in dynamic constrained search space. The proposed algorithm involves two phases, namely the CB phase and the SS phase. In the CB phase, the mutual information is utilized as the restriction to limit the search space, and a binding parameter is introduced to the novel encoding scheme so that the search space can be dynamically changed in the evolutionary process. In the SS phase, a new operator is developed to pass on the excellent genes from generation to generation, and an update principle for the binding parameter is exploited for the dynamic selection of the search space. We conduct the comparative experiments on the benchmark network data sets and provide performance and applicability analysis of our proposed method. The experimental results show that the new algorithm is effective in learning the BN structures.

A dynamic search space strategy for swarm intelligence

As an appendix which is designed to embed in one of complete swarm intelligence algorithms, a novel strategy named dynamic-search-spaces (DS) is proposed to deal with the premature convergence of those algorithms. For realising the decrement of search space, the differences or the distances between individuals and the global performance are to form a threshold for building the self-adaption system. When the value calculating the rate of those individuals sitting near the global performance reaches a stated percentage, the system is working to readjust the borders of search space by the site of the global performance. The search space will be compressed to close the global performance as the centre. After each readjustment, the re-initialisation to distribute individuals in the whole search space should be achieved to enhance individuals' vitality which moves away from the premature convergence and improves the performance of each individual. Meanwhile, the simpler verifications are provided. The improvements of results are exhibited embedding in the genetic algorithm, the particle swarm optimisation and the differential evolution. Thus, this dynamic search space scheme can be embedded in most swarm intelligence algorithms easily.

A dynamic search space strategy for swarm intelligence

As an appendix which is designed to embed in one of complete swarm intelligence algorithms, a novel strategy named dynamic-search-spaces (DS) is proposed to deal with the premature convergence of those algorithms. For realising the decrement of search space, the differences or the distances between individuals and the global performance are to form a threshold for building the self-adaption system. When the value calculating the rate of those individuals sitting near the global performance reaches a stated percentage, the system is working to readjust the borders of search space by the site of the global performance. The search space will be compressed to close the global performance as the centre. After each readjustment, the re-initialisation to distribute individuals in the whole search space should be achieved to enhance individuals' vitality which moves away from the premature convergence and improves the performance of each individual. Meanwhile, the simpler verifications are provided. The improvements of results are exhibited embedding in the genetic algorithm, the particle swarm optimisation and the differential evolution. Thus, this dynamic search space scheme can be embedded in most swarm intelligence algorithms easily.

R=1/2 Self-Doubly 조직 직교 길쌈부호를 찾는 효율적인 최적 스팬 알고리듬

In this paper, a new method for finding optimal and short span in Convolutional Self-Doubly Orthogonal(CDO) codes are proposed. This new algorithm based on Parallel Implicitly-Exhaustive search, where we applied dynamic search space reduction methods in order to reduce computational time for finding Optimal Span for R=1/2 rate CDO codes. The simulation results shows that speedup and error correction performance of the new algorithm is better.

Identification of Predictive Cis-Regulatory Elements Using a Discriminative Objective Function and a Dynamic Search Space.

The generation of genomic binding or accessibility data from massively parallel sequencing technologies such as ChIP-seq and DNase-seq continues to accelerate. Yet state-of-the-art computational approaches for the identification of DNA binding motifs often yield motifs of weak predictive power. Here we present a novel computational algorithm called MotifSpec, designed to find predictive motifs, in contrast to over-represented sequence elements. The key distinguishing feature of this algorithm is that it uses a dynamic search space and a learned threshold to find discriminative motifs in combination with the modeling of motifs using a full PWM (position weight matrix) rather than k-mer words or regular expressions. We demonstrate that our approach finds motifs corresponding to known binding specificities in several mammalian ChIP-seq datasets, and that our PWMs classify the ChIP-seq signals with accuracy comparable to, or marginally better than motifs from the best existing algorithms. In other datasets, our algorithm identifies novel motifs where other methods fail. Finally, we apply this algorithm to detect motifs from expression datasets in C. elegans using a dynamic expression similarity metric rather than fixed expression clusters, and find novel predictive motifs.