Traditional Crossover Operators Research Articles

A surrogate evolutionary neural architecture search algorithm for graph neural networks

Due to the unique construction module and design of graph neural networks (GNNs), neural architecture search (NAS) methods specifically for GNNs have become a promising research hotspot in recent years. Among the existing methods, one class of methods microscopically searches for the constituent components of network layers. However, most of them ignore the topology connections between network layers or the feature fusion strategies. Another class of methods, called differentiable architecture search methods, has the advantage of searching topology connections and feature fusion strategies. However, constrained by the requirement of predefining all candidate operations, these methods can only sample a limited number of network layers. In this paper, we propose a surrogate evolutionary graph neural architecture search (GNAS) algorithm whose search space contains not only the microscopic network layer components but also topology connections and feature fusion strategies (called CTFGNAS). The GNN sampled in CTFGNAS is represented by a simple one-dimensional vector and does not fix the network depth. To address the problem that traditional crossover and mutation operators applied to GNAS may produce illegal solutions, we design a repair operation to guarantee the legitimacy of the solutions. The network depth is also increased with a large probability in the mutation operation to alleviate the oversmoothing problem. In addition, to cope with the challenge of computational resources due to the increased search space, we form a surrogate model with three classical regression models, where only a small number of solutions are truly evaluated for their fitness, and the remaining large number of solutions are predicted for their fitness by the surrogate model. Finally, experiments are executed on six widely used real-world datasets. The experimental results illustrate that CTFGNAS obtains more effective results than the state-of-the-art handcrafted GNNs and GNAS methods on all datasets. CTFGNAS is now available on the following website: https://github.com/chnyliu/CTFGNAS.

An Evolutionary Algorithm with Gene Ontology-Aware Crossover Operator for Protein Complex Detection

Evolutionary algorithms (EAs), as global search methods, are proved to be more robust than their counterpart local heuristics for detecting protein complexes in protein-protein interaction (PPI) networks. Typically, the source of robustness of these EAs comes from their components and parameters. These components are solution representation, selection, crossover, and mutation. Unfortunately, almost all EA based complex detection methods suggested in the literature were designed with only canonical or traditional components. Further, topological structure of the protein network is the main information that is used in the design of almost all such components. The main contribution of this paper is to formulate a more robust EA with more biological consistency. For this purpose, a new crossover operator is suggested where biological information in terms of both gene semantic similarity and protein functional similarity is fed into its design. To reflect the heuristic roles of both semantic and functional similarities, this paper introduces two gene ontology (GO) aware crossover operators. These are direct annotation-aware and inherited annotation-aware crossover operators. The first strategy is handled with the direct gene ontology annotation of the proteins, while the second strategy is handled with the directed acyclic graph (DAG) of each gene ontology term in the gene product. To conduct our experiments, the proposed EAs with GO-aware crossover operators are compared against the state-of-the-art heuristic, canonical EAs with the traditional crossover operator, and GO-based EAs. Simulation results are evaluated in terms of recall, precision, and F measure at both complex level and protein level. The results prove that the new EA design encourages a more reliable treatment of exploration and exploitation and, thus, improves the detection ability for more accurate protein complex structures.

Dynastic Potential Crossover Operator.

An optimal recombination operator for two-parent solutions provides the best solution among those that take the value for each variable from one of the parents (gene transmission property). If the solutions are bit strings, the offspring of an optimal recombination operator is optimal in the smallest hyperplane containing the two parent solutions. Exploring this hyperplane is computationally costly, in general, requiring exponential time in the worst case. However, when the variable interaction graph of the objective function is sparse, exploration can be done in polynomial time. In this article, we present a recombination operator, called Dynastic Potential Crossover (DPX), that runs in polynomial time and behaves like an optimal recombination operator for low-epistasis combinatorial problems. We compare this operator, both theoretically and experimentally, with traditional crossover operators, like uniform crossover and network crossover, and with two recently defined efficient recombination operators: partition crossover and articulation points partition crossover. The empirical comparison uses NKQ Landscapes and MAX-SAT instances. DPX outperforms the other crossover operators in terms of quality of the offspring and provides better results included in a trajectory and a population-based metaheuristic, but it requires more time and memory to compute the offspring.

Multi-objective workflow scheduling based on genetic algorithm in cloud environment

In recent years, cloud computing plays a crucial role in many real applications. Thus, how to solve workflow scheduling problems, i.e., allocating and scheduling different resources, under the cloud computing environment becomes more important. Although some evolutionary algorithms (EAs) can solve workflow scheduling problems with a small scale, they show some disadvantages on larger scale workflow applications. In this paper, a multi-objective genetic algorithm (MOGA) is applied to optimize workflow scheduling problems. To enhance the search efficiency, this study proposes an initialization scheduling sequence scheme, in which each task’s data size is considered when initializing its virtual machine (VM) instance. Relying on the initial scheduling sequence, a proper trade-off between the makespan and the energy consumption, which are two optimization objectives in this study, can be achieved. In the early evolution stage, traditional crossover and mutate operators are performed to keep the population’s exploration. On the contrary, the longest common subsequence (LCS) of multiple elite individuals, which can be regarded as a favorable gene block, is saved during the later evolution stage. Based on the LCS, the probability of some favorable gene blocks being destroyed will be reduced when performing the crossover operator and the mutate operator. Hence, the integration of the LCS in GA can satisfy different requirements in different evolution stages, and then to attain a balance between the exploration and the exploitation. Extensive experimental results verify that the proposed GA combined with LCS, named as GALCS in this paper, can find a better Pareto front than the ordinary GA as well as other state-of-the-art algorithms. Furthermore, effectivenesses of the new proposed strategies are also verified by a set of experiments.

Evolution test by improved genetic algorithm with application to performance limit evaluation of automatic parallel parking system

Performance limit evaluation of automatic driving system before putting into the market is critical for driving safety. The evolution test by genetic algorithm (GA) is a method by iteratively generating new test scenarios according to the last test results. To avoid its blind search for better efficiency, a scenario complexity index is proposed to measure the test effectiveness indirectly and guide the evolution process under the assumption that a complex scenario is more challenging to realise automatic driving. The traditional crossover and mutation operators are modified to generate more complex scenarios to improve the test efficiency. The advantage of the improved crossover/mutation operators in increasing the offspring's scenario complexity index is analysed in theory. Moreover, the influence of the design parameters on the evolution test process and the global convergence are also discussed. The new evolution test by this improved GA has been applied to find the collision condition of a parallel automatic parking system to validate its effectiveness.

The Application of an Adaptive Genetic Algorithm Based on Collision Detection in Path Planning of Mobile Robots

An adaptive genetic algorithm based on collision detection (AGACD) is proposed to solve the problems of the basic genetic algorithm in the field of path planning, such as low convergence path quality, many iterations required for convergence, and easily falling into the local optimal solution. First, this paper introduces the Delphi weight method to evaluate the weight of path length, path smoothness, and path safety in the fitness function, and a collision detection method is proposed to detect whether the planned path collides with obstacles. Then, the population initialization process is improved to reduce the program running time. After comprehensively considering the population diversity and the number of algorithm iterations, the traditional crossover operator and mutation operator are improved, and the adaptive crossover operator and adaptive mutation operator are proposed to avoid the local optimal solution. Finally, an optimization operator is proposed to improve the quality of convergent individuals through the second optimization of convergent individuals. The simulation results show that the adaptive genetic algorithm based on collision detection is not only suitable for simulation maps with various sizes and obstacle distributions but also has excellent performance, such as greatly reducing the running time of the algorithm program, and the adaptive genetic algorithm based on collision detection can effectively solve the problems of the basic genetic algorithm.

Reward-based epigenetic learning algorithm for a decentralised multi-agent system

PurposeThis paper aims to propose a novel epigenetic learning (EpiLearn) algorithm, which is designed specifically for a decentralised multi-agent system such as swarm robotics.Design/methodology/approachFirst, this paper begins with overview of swarm robotics and the challenges in designing swarm behaviour automatically. This should indicate the direction of improvements required to enhance an automatic swarm design. Second, the evolutionary learning (EpiLearn) algorithm for a swarm system using an epigenetic layer is formulated and discussed. The algorithm is then tested through various test functions to investigate its performance. Finally, the results are discussed along with possible future research directions.FindingsThrough various test functions, the algorithm can solve non-local and many local minima problems. This article also shows that by using a reward system, the algorithm can handle the deceptive problem which often occurs in dynamic problems. Moreover, utilization of rewards from the environment in the form of a methylation process on the epigenetic layer improves the performance of traditional evolutionary algorithms applied to automatic swarm design. Finally, this article shows that a regeneration process that embeds an epigenetic layer in the inheritance process performs better than a traditional crossover operator in a swarm system.Originality/valueThis paper proposes a novel method for automatic swarm design by taking into account the importance of multi-agent settings and environmental characteristics surrounding the swarm. The novel evolutionary learning (EpiLearn) algorithm using an epigenetic layer gives the swarm the ability to perform co-evolution and co-learning.

Delay-Bounded and Cost-Limited RSU Deployment in Urban Vehicular Ad Hoc Networks.

As an auxiliary facility, roadside units (RSUs) can well improve the shortcomings incurred by ad hoc networks and promote network performance in a vehicular ad hoc network (VANET). However, deploying a large number of RSUs will lead to high installation and maintenance costs. Therefore, trying to find the best locations is a key issue when deploying RSUs with the set delay and budget. In this paper, we study the delay-bounded and cost-limited RSU deployment (DBCL) problem in urban VANET. We prove it is non-deterministic polynomial-time hard (NP-hard), and a binary differential evolution scheme is proposed to maximize the number of roads covered by deploying RSUs. Opposite-based learning is introduced to initialize the first generation, and a binary differential mutation operator is designed to obtain binary coding. A random variable is added to the traditional crossover operator to increase population diversity. Also, a greedy-based individual reparation and promotion algorithm is adopted to repair infeasible solutions violating given constraints, and to gain optimal feasible solutions with the compromise of given limits. Moreover, after selection, a solution promotion algorithm is executed to promote the best solution found in generation. Simulation is performed on analog trajectories sets, and results show that our proposed algorithm has a higher road coverage ratio and lower packet loss compared with other schemes.

유전자 알고리즘의 수렴 속도 향상을 통한 효과적인 로봇 길 찾기 알고리즘

유전자 알고리즘은 초기 해 집합을 대상으로 해 집합의 평가와 유전자 연산자의 적용, 자연 선택 등의 과정을 반복하여 최적 해를 찾는 탐색 알고리즘이다. 유전자 알고리즘을 설계할 때 사용한 선택 전략, 세대교체 방법, 유전자 연산자 등은 유전자 알고리즘의 탐색 효율성에 영향을 준다. 본 논문에서는 시간 제약이 있는 상황에서의 로봇 경로 탐색을 위해 기존의 유전자 알고리즘보다 빠르게 수렴하는 유전자 알고리즘을 제안한다. 로봇 경로 탐색 시 긴급한 상황에서 유전자 알고리즘은 연산을 위한 충분한 시간을 확보하지 못 하게 되고, 이는 최종적으로 찾아낸 경로의 질을 떨어뜨린다. 제안하는 알고리즘은 빠른 수렴을 위한 선택 전략, 세대교체 방법을 사용하였으며, 유전자 연산자로는 전통적인 교차, 돌연변이 외에 경로의 길이를 줄이기 위한 단축 연산자를 추가로 사용하였다. 이를 통해 제안하는 알고리즘은 적은 세대 수에도 빠르게 짧은 경로를 찾아낸다. The Genetic algorithm is a search algorithm using evaluation, genetic operator, natural selection to populational solution iteratively. The convergence and divergence characteristic of genetic algorithm are affected by selection strategy, generation replacement method, genetic operator when genetic algorithm is designed. This paper proposes fast convergence genetic algorithm for time-limited robot path planning. In urgent situation, genetic algorithm for robot path planning does not have enough time for computation, resulting in quality degradation of found path. Proposed genetic algorithm uses fast converging selection strategy and generation replacement method. Proposed genetic algorithm also uses not only traditional crossover and mutation operator but additional genetic operator for shortening the distance of found path. In this way, proposed genetic algorithm find reasonable path in time-limited situation.

A multi-parent genetic algorithm for the quadratic assignment problem

Instead of using traditional (two-parent) crossover operator, multi-parent crossover operator is used in genetic algorithms to improve solution quality for many numerical optimization problems. However, very few literatures are available on multi-parent crossover operator for combinatorial optimization problems, especially, quadratic assignment problem (QAP). This paper proposes a multi-parent extension of the sequential constructive crossover (MPSCX), which is a generalization of the traditional sequential constructive crossover (SCX), for the QAP. Then a multi-parent genetic algorithm (MPGA) using MPSCX is developed. Experimental results on ten QAPLIB instances show that our MPGA significantly improves GA using SCX by up to 1.75 % in average assignment cost with maximum of 2.79 % away from the best known solution value. Finally, the efficiency of our MPGA is compared against MPGA using an existing multi-parent crossover for the problem. Experimental results show that our MPGA is better.

Optimisation algorithm for logistics distribution route based on Prufer codes

For logistics distribution route problem, this paper presents an improved genetic algorithm, combined with the thought of tree coded, to solve single vehicle logistics distribution route problem. The algorithm adopted Prufer array coding reducing three units length of chromosome, and the decoding method did not increase time complexity of chromosome decoded to calculate objective function value when compared with the traditional coding method which coded chromosome according to sequence traversing each node. In order to prevent the feasibility of solution being destructed by the traditional crossover and mutation operators, a single transposition operator and an ortho-position exchanging operator were also designed. The following cases show that the algorithm has a higher solving efficiency and is a great help to solve large-scale route optimisation problem.

An improved NSGA-II algorithm for multi-objective lot-streaming flow shop scheduling problem

Crossover and mutation operators in NSGA-II are random and aimless, and encounter difficulties in generating offspring with high quality. Aiming to overcoming these drawbacks, we proposed an improved NSGA-II algorithm (INSGA-II) and applied it to solve the lot-streaming flow shop scheduling problem with four criteria. We first presented four variants of NEH heuristic to generate the initial population, and then incorporated the estimation of distribution algorithm and a mutation operator based on insertion and swap into NSGA-II to replace traditional crossover and mutation operators. Last but not least, we performed a simple and efficient restarting strategy on the population when the diversity of the population is smaller than a given threshold. We conducted a serial of experiments, and the experimental results demonstrate that the proposed algorithm outperforms the comparative algorithms.

An Improved Genetic Algorithm for the Distributed and Flexible Job-shop Scheduling problem

The Distributed and Flexible Job-shop Scheduling problem (DFJS) considers the scheduling of distributed manufacturing environments, where jobs are processed by a system of several Flexible Manufacturing Units (FMUs). Distributed scheduling problems deal with the assignment of jobs to FMUs and with determining the scheduling of each FMU, in terms of assignment of each job operation to one of the machines able to work it (job-routing flexibility) and sequence of operations on each machine. The objective is to minimize the global makespan over all the FMUs. This paper proposes an Improved Genetic Algorithm to solve the Distributed and Flexible Job-shop Scheduling problem. With respect to the solution representation for non-distributed job-shop scheduling, gene encoding is extended to include information on job-to-FMU assignment, and a greedy decoding procedure exploits flexibility and determines the job routings. Besides traditional crossover and mutation operators, a new local search based operator is used to improve available solutions by refining the most promising individuals of each generation. The proposed approach has been compared with other algorithms for distributed scheduling and evaluated with satisfactory results on a large set of distributed-and-flexible scheduling problems derived from classical job-shop scheduling benchmarks.

Using Genetic Programming for an Advanced Performance Assessment of Industrially Relevant Heterogeneous Catalysts

Beside the ease and speed brought by automated synthesis stations and reactors technologies in materials science, adapted informatics tools must be further developed in order to handle the increase of throughput and data volume, and not to slow down the whole process. This article reports the use of genetic programming (GP) in heterogeneous catalysis. Despite the fact that GP has received only little attention in this domain, it is shown how such an approach can be turned into a very singular and powerful tool for solid optimization, discovery, and monitoring. Jointly with neural networks, the GP paradigm is employed in order to accurately and automatically estimate the whole curve “conversion vs. time” in the epoxidation of large olefins using titanosilicates, Ti-MCM-41 and Ti-ITQ-2, as catalysts. In contrast to previous studies in combinatorial materials science and high-throughput screening, it was possible to estimate the entire evolution of the catalytic reaction for unsynthesized catalysts. Consequently, the evaluation of the performance of virtual solids is not reduced to a single point (e.g., the conversion level at only one given reaction time or the initial reaction rate). The methodology is thoroughly detailed, while stressing on the comparison between the recently proposed Context Aware Crossover (CAX) and the traditional crossover operator.

Virus coevolution partheno-genetic algorithms for optimal sensor placement

A virus coevolutionary partheno-genetic algorithm (VEPGA), which combined a partheno-genetic algorithm (PGA) with virus evolutionary theory, is proposed to place sensors optimally on a large space structure for the purpose of modal identification. The VEPGA is composed of a host population of candidate solutions and a virus population of substrings of host individuals. The traditional crossover and mutation operators in genetic algorithm are repealed and their functions are implemented by particular partheno-genetic operators which are suitable to combinatorial optimization problems. Three different optimal sensor placement performance index, one aim on the maximization of linear independence, one aim on the maximization of modal energy and the last is a combination of the front two indices, have been investigated. The algorithm is applied to two examples: sensor placement for a portal frame and a concrete arc dam. Results show that the proposed VEPGA outperforms the sequential reduction procedure (SRP) and PGA. The combined performance index makes an excellent compromise between the linear independence aimed index and the modal energy aimed index.

Memetic algorithms for the unconstrained binary quadratic programming problem

This paper presents a memetic algorithm, a highly effective evolutionary algorithm incorporating local search for solving the unconstrained binary quadratic programming problem (BQP). To justify the approach, a fitness landscape analysis is conducted experimentally for several instances of the BQP. The results of the analysis show that recombination-based variation operators are well suited for the evolutionary algorithms with local search. Therefore, the proposed approach includes — besides a highly effective randomized k-opt local search — a new variation operator that has been tailored specially for the application in the hybrid evolutionary framework. The operator is called innovative variation and is fundamentally different from traditional crossover operators, since new genetic material is included in the offspring which is not contained in one of the parents. The evolutionary heuristic is tested on 35 publicly available BQP instances, and it is shown experimentally that the algorithm is capable of finding best-known solutions to large BQPs in a short time and with a high frequency. In comparison to other approaches for the BQP, the approach appears to be much more effective, particularly for large instances of 1000 or 2500 binary variables.

Genetic algorithm with pigeon-hole coding scheme for solving sequencing problems

One ofthe major difficulties in applying genetic algorithms in sequencing problems is the production of infeasible sequences by crossover operations. In this article we propose a new coding scheme based on a pigeon-hole idea.The Pigeon-hole coding scheme can guarantee the feasibility ofthe offspring generated by crossover operations. In addition,we adopta Randomized Single-point-Crossover operator and a Modified-Mutation operator in the algorithm,which are more effective as compared with the traditional crossover and mutation operators. Our pigeon-hole-based genetic algorithm is successfully applied in solving two NP-hard sequencing problems.

Genetic beam search for gate matrix layout

The paper presents a novel implementation of the genetic algorithm in combination with beam search for gate matrix layout. This is a permutation problem, for which the traditional genetic crossover operator results in repetition of gates, and therefore the GA is not applicable without modification. However, the GA is a very efficient stochastic optimisation technique, and is easily parallelisable. To adapt it to the gate matrix layout problem, the principles of beam search have been used. The gates are ranked according to their connectivity with each other, and the ranks of the gates to be placed next to each other are picked by the GA. A beam value is used to restrict the ranks of the gates placed next to each other to a low value. This reduces the search space to be explored, and thus results in a more efficient search. The algorithm produces better results compared to a graph-theoretic approach on published netlists.