Extremal optimization

About: Extremal optimization is a research topic. Over the lifetime, 1168 publications have been published within this topic receiving 104943 citations.

Graph neural network inspired algorithm for unsupervised network community detection

Abstract: Abstract Network community detection often relies on optimizing partition quality functions, like modularity. This optimization appears to be a complex problem traditionally relying on discrete heuristics. And although the problem could be reformulated as continuous optimization, direct application of the standard optimization methods has limited efficiency in overcoming the numerous local extrema. However, the rise of deep learning and its applications to graphs offers new opportunities. And while graph neural networks have been used for supervised and unsupervised learning on networks, their application to modularity optimization has not been explored yet. This paper proposes a new variant of the recurrent graph neural network algorithm for unsupervised network community detection through modularity optimization. The new algorithm’s performance is compared against the state-of-the-art methods. The approach also serves as a proof-of-concept for the broader application of recurrent graph neural networks to unsupervised network optimization.

Simulation Based Multi-objective Extremal Optimization Algorithm for Electronic Reconnaissance Satellites Scheduling Problem

Abstract: A multi-objective chance constrained programming model(MOCCPM) for electronic reconnaissance satellites scheduling problem(ERSSP) is presented. MOCCPM takes the uncertainties in the course of satellite electronic reconnaissance into account, as well as the capabilities and usage restrictions of the electronic reconnaissance satellites. Then a Monte Carlo simulation based multi-objective extremal optimization (MCSBMOEO) algorithm is proposed. Penalty function based fitness assignment ensures the efficient evolution. Problem specific mutation operator ensures the feasibility of the offspring so as to prevent the algorithm from falling into local optimum. External archive is to keep the non-dominated solutions and guarantee their diversity. Monte Carlo sampling is to address the stochastic nature of ERSSP. The experiment results testified that the algorithm can solve ERSSP effectively.

A population extremal optimization based modified constrained generalized predictive control method

Abstract: As one of the most popular and successful methods in industrial applications, model predictive control (MPC) has attracted increasing interest in the past two decades. However, one of open issues in this research filed is how to solve the constrained nonlinear optimization problems in MPC. From the perspective of evolutionary algorithm, this paper presents a novel population extremal optimization (PEO) based modified constrained generalized predictive control (CGPC) method called CGPC-PEO. The key idea behind the proposed CGPC-PEO is using PEO for rolling optimization to minimize the weighted objective function subjecting to a set of constraints. Its superiority to other evolutionary algorithms such as genetic algorithm and particle swarm optimization based CGPC is demonstrated by the simulation results on an industrial process plant.

Application of ant colony algorithm in reliability optimization of complex system

Abstract: Reliability optimization of complex system is a typical NP-hard problem.Without taking into account the system's specific connecting form,by regarding every component as one level and the generated random data as network nodes,ant colony optimization(ACO) was successfully adopted to the mentioned optimization problem to search the optimal solution which could not be obtained by other algorithms.The simulation results indicate that ACO can find the optimal solution more quickly.Like other heuristic algorithms,ACO can also overcome the problem of computation complexity of combination optimization effectively.

The application of ant colony algorithmin the continuous space optimization problems

Abstract: In this paper, a new type of ant colony algorithm's explorative strategy which can be applied to solve the continuous space optimization problems was studied. This new strategy could promote the efficiency, the diversity and the stochasty of exploration process, and it can also avoid the restrictions about the optimization problem's functions whether they are continuous or differential. A feasible way is presented for applying the ant colony algorithm to practice. The numerical results demonstrates that the approximate optimization result for whole domain can be available efficiently.