Jinghui Zhong

Evolutionary Computation Meets Machine Learning: A Survey

Evolutionary computation (EC) is a kind of optimization methodology inspired by the mechanisms of biological evolution and behaviors of living organisms. In the literature, the terminology evolutionary algorithms is frequently treated the same as EC. This article focuses on making a survey of researches based on using ML techniques to enhance EC algorithms. In the framework of an ML-technique enhanced-EC algorithm (MLEC), the main idea is that the EC algorithm has stored ample data about the search space, problem features, and population information during the iterative search process, thus the ML technique is helpful in analyzing these data for enhancing the search performance. The paper presents a survey of five categories: ML for population initialization, ML for fitness evaluation and selection, ML for population reproduction and variation, ML for algorithm adaptation, and ML for local search.

Implementation of an Ant Colony Optimization technique for job shop scheduling problem

Research on optimization of the job shop scheduling problem (JSP) is one of the most significant and promising areas of optimization. Instead of the traditional optimization method, this paper presents an investigation into the use of an Ant Colony System (ACS) to optimize the JSP. The main characteristics of this system are positive feedback, distributed computation, robustness and the use of a constructive greedy heuristic. In this paper, an improvement of the performance of ACS will be discussed. The numerical experiments of ACS were implemented in a small JSP. The optimized results of the ACS are favourably compared with the traditional optimization methods.

A Differential Evolution Algorithm With Dual Populations for Solving Periodic Railway Timetable Scheduling Problem

Railway timetable scheduling is a fundamental operational problem in the railway industry and has significant influence on the quality of service provided by the transport system. This paper explores the periodic railway timetable scheduling (PRTS) problem, with the objective to minimize the average waiting time of the transfer passengers. Unlike traditional PRTS models that only involve service lines with fixed cycles, this paper presents a more flexible model by allowing the cycle of service lines and the number of transfer passengers to vary with the time period. An enhanced differential evolution (DE) algorithm with dual populations, termed “dual-population DE” (DP-DE), was developed to solve the PRTS problem, yielding high-quality solutions. In the DP-DE, two populations cooperate during the evolution; the first focuses on global search by adopting parameter settings and operators that help maintain population diversity, while the second one focuses on speeding up convergence by adopting parameter settings and operators that are good for local fine tuning. A novel bidirectional migration operator is proposed to share the search experience between the two populations. The proposed DP-DE has been applied to optimize the timetable of the Guangzhou Metro system in Mainland China and six artificial periodic railway systems. Two conventional deterministic algorithms and seven highly regarded evolutionary algorithms are used for comparison. The comparison results reveal that the performance of DP-PE is very promising.

Self-Learning Gene Expression Programming

In this paper, a novel self-learning gene expression programming (GEP) methodology named SL-GEP is proposed to improve the search accuracy and efficiency of GEP. In contrast to the existing GEP variants, the proposed SL-GEP features a novel chromosome representation in which each chromosome is embedded with subfunctions that can be deployed to construct the final solution. As part of the chromosome, the subfunctions are self-learned or self-evolved by the proposed algorithm during the evolutionary search. By encompassing subfunctions or any partial solution as input arguments of another subfunction, the proposed SL-GEP facilitates the formation of sophisticated, higher-order, and constructive subfunctions that improve the accuracy and efficiency of the search. Further, a novel search mechanism based on differential evolution is proposed for the evolution of chromosomes in the SL-GEP. The proposed SL-GEP is simple, generic and has much fewer control parameters than the traditional GEP variants. The proposed SL-GEP is validated on 15 symbolic regression problems and six even-parity problems. Experimental results show that the proposed SL-GEP offers enhanced performances over several state-of-the-art algorithms in terms of accuracy and search efficiency.

Density-based evolutionary framework for crowd model calibration

Crowd modeling and simulation is an important and active research field, with a wide range of applications such as computer games, military training and evacuation modeling. One important issue in crowd modeling is model calibration through parameter tuning, so as to produce desired crowd behaviors. Common methods such as trial-and-error are time consuming and tedious. This paper proposes an evolutionary framework to automate the crowd model calibration process. In the proposed framework, a density-based matching scheme is introduced. By using the dynamic density of the crowd over time, and a weight landscape to emphasize important spatial regions, the proposed matching scheme provides a generally applicable way to evaluate the simulated crowd behaviors. Besides, a hybrid search mechanism based on differential evolution is proposed to efficiently tune parameters of crowd models. Simulation results demonstrate that the proposed framework is effective and efficient to calibrate the crowd models in order to produce desired macroscopic crowd behaviors.

Ant colony optimization algorithm for lifetime maximization in wireless sensor network with mobile sink

In wireless sensor networks (WSNs), sensors near the sink can be burdened with a large amount of traffic, because they have to transmit data generated by themselves and those far away from the sink. Hence the sensors near the sink would deplete their energy much faster than the others, which results in a short network lifetime. Using mobile sink is an effective way to tackle this issue. This paper explores the problem of determining the optimal movements of the mobile sink to maximize the network lifetime. A novel ant colony optimization algorithm (ACO), namely the ACO-MSS, is developed to solve the problem. The proposed ACO-MSS takes advantage of the global search ability of ACO and adopts effective heuristic information to find a near globally optimal solution. Multiple practical factors such as the forbidden regions and the maximum moving distance of the sink are taken into account to facilitate the real applications. The proposed ACO-MSS is validated by a series of simulations on WSNs with different characteristics. The simulation results demonstrate the effectiveness of the proposed algorithms.

Energy-efficient local wake-up scheduling in wireless sensor networks

Scheduling sensor activities is an effective way to prolong the lifetime of wireless sensor networks (WSNs). In this paper, we explore the problem of wake-up scheduling in WSNs where sensors have different lifetime. A novel local wake-up scheduling (LWS) strategy is proposed to prolong the network lifetime with full coverage constraint. In the LWS strategy, sensors are divided into a first layer set and a successor set. The first layer set which satisfies the coverage constraint is activated at the beginning. Once an active sensor runs out of energy, some sensors in the successor set will be activated to satisfy the coverage constraint. Based on the LWS strategy, this paper presents an ant colony optimization based method, namely mc-ACO, to maximize the network lifetime. The mc-ACO is validated by performing simulations on WSNs with different characteristics. A recently published genetic algorithm based wake-up scheduling method and a greedy based method are used for comparison. Simulation results reveal that mc-ACO yields better performance than the two algorithms.

Adaptive crossover and mutation in genetic algorithms based on clustering technique

Instead of having fixed px and pm, this paper presents the use of fuzzy logic to adaptively tune px and pm for optimization of power electronic circuits throughout the process. By applying the K-means algorithm, distribution of the population in the search space is clustered in each training generation. Inferences of px and pm are performed by a fuzzy-based system that fuzzifies the relative sizes of the clusters containing the best and worst chromosomes. The proposed adaptation method is applied to optimize a buck regulator that requires satisfying some static and dynamic requirements. The optimized circuit component values, the regulators performance, and the convergence rate in the training are favorably compared with the GAs using fixed px and p.

MP-EDA: a robust estimation of distribution algorithm with multiple probabilistic models for global continuous optimization

Extending Estimation of distribution algorithms (EDAs) to the continuous field is a promising and challenging task. With a single probabilistic model, most existing continuous EDAs usually suffer from the local stagnation or a low convergence speed. This paper presents an enhanced continuous EDA with multiple probabilistic models (MP-EDA). In the MP-EDA, the population is divided into two subpopulations. The one involved by histogram model is used to roughly capture the global optima, whereas the other involved by Gaussian model is aimed at finding highly accurate solutions. During the evolution, a migration operation is periodically carried out to exchange some best individuals of the two subpopulations. Besides, the MP-EDA adaptively adjusts the offspring size of each subpopulation to improve the searching efficiency. The effectiveness of the MP-EDA is investigated by testing ten benchmark functions. Compared with several state-of-the-art evolutionary computations, the proposed algorithm can obtain better results in most test cases.

Ea-based evacuation planning using agent-based crowd simulation

Safety planning for crowd evacuation is an important and active research topic nowadays. One important issue is to devise the evacuation plans of individuals in emergency situations so as to reduce the total evacuation time. This paper proposes a novel evolutionary algorithm (EA)-based methodology, together with agent-based crowd simulation, to solve the evacuation planning problem. The proposed method features a novel segmentation strategy which divides the entire evacuation region into sub-regions based on a discriminant function. Each sub-region is assigned with an exit gate, and individuals in a sub-region will run toward the corresponding exit gate for evacuation. In this way, the evacuation planning problem is converted to a symbolic regression problem. Then an evolutionary algorithm, using agent-based crowd simulation as fitness function, is developed to search for the global optimal solution. The simulation results on different scenarios demonstrate that the proposed method is effective to reduce the evacuation time.