The problem of potential field inversion can be become that of solving system of linear equations by using of linear processing. There are a lot of algorithms for solving any system of linear equations, and the regula...The problem of potential field inversion can be become that of solving system of linear equations by using of linear processing. There are a lot of algorithms for solving any system of linear equations, and the regularized method is one of the best algorithms. But there is a shortcoming in application with the regularized method, viz. the optimum regularized parameter must be determined by experience, so it is difficulty to obtain an optimum solution. In this paper, an iterative algorithm for solving any system of linear equations is discussed, and a sufficient and necessary condition of the algorithm convergence is presented and proved. The algorithm is convergent for any starting point, and the optimum solution can be obtained, in particular, there is no need to calculate the inverse matrix in the algorithm. The typical practical example shows the iterative algorithm is simple and practicable, and the inversion effect is better than that of regularized method.展开更多
Cooperative path planning is an important area in fixed-wing UAV swarm.However,avoiding multiple timevarying obstacles and avoiding local optimum are two challenges for existing approaches in a dynamic environment.Fir...Cooperative path planning is an important area in fixed-wing UAV swarm.However,avoiding multiple timevarying obstacles and avoiding local optimum are two challenges for existing approaches in a dynamic environment.Firstly,a normalized artificial potential field optimization is proposed by reconstructing a novel function with anisotropy in each dimension,which can make the flight speed of a fixed UAV swarm independent of the repulsive/attractive gain coefficient and avoid trapping into local optimization and local oscillation.Then,taking into account minimum velocity and turning angular velocity of fixed-wing UAV swarm,a strategy of decomposing target vector to avoid moving obstacles and pop-up threats is proposed.Finally,several simulations are carried out to illustrate superiority and effectiveness.展开更多
A novel method was designed to solve reinforcement learning problems with artificial potential field.Firstly a reinforcement learning problem was transferred to a path planning problem by using artificial potential fi...A novel method was designed to solve reinforcement learning problems with artificial potential field.Firstly a reinforcement learning problem was transferred to a path planning problem by using artificial potential field(APF),which was a very appropriate method to model a reinforcement learning problem.Secondly,a new APF algorithm was proposed to overcome the local minimum problem in the potential field methods with a virtual water-flow concept.The performance of this new method was tested by a gridworld problem named as key and door maze.The experimental results show that within 45 trials,good and deterministic policies are found in almost all simulations.In comparison with WIERING's HQ-learning system which needs 20 000 trials for stable solution,the proposed new method can obtain optimal and stable policy far more quickly than HQ-learning.Therefore,the new method is simple and effective to give an optimal solution to the reinforcement learning problem.展开更多
Region partition(RP) is the key technique to the finite element parallel computing(FEPC),and its performance has a decisive influence on the entire process of analysis and computation.The performance evaluation index ...Region partition(RP) is the key technique to the finite element parallel computing(FEPC),and its performance has a decisive influence on the entire process of analysis and computation.The performance evaluation index of RP method for the three-dimensional finite element model(FEM) has been given.By taking the electric field of aluminum reduction cell(ARC) as the research object,the performance of two classical RP methods,which are Al-NASRA and NGUYEN partition(ANP) algorithm and the multi-level partition(MLP) method,has been analyzed and compared.The comparison results indicate a sound performance of ANP algorithm,but to large-scale models,the computing time of ANP algorithm increases notably.This is because the ANP algorithm determines only one node based on the minimum weight and just adds the elements connected to the node into the sub-region during each iteration.To obtain the satisfied speed and the precision,an improved dynamic self-adaptive ANP(DSA-ANP) algorithm has been proposed.With consideration of model scale,complexity and sub-RP stage,the improved algorithm adaptively determines the number of nodes and selects those nodes with small enough weight,and then dynamically adds these connected elements.The proposed algorithm has been applied to the finite element analysis(FEA) of the electric field simulation of ARC.Compared with the traditional ANP algorithm,the computational efficiency of the proposed algorithm has been shortened approximately from 260 s to 13 s.This proves the superiority of the improved algorithm on computing time performance.展开更多
The self-potential method is widely used in environmental and engineering geophysics. Four intelligent optimization algorithms are adopted to design the inversion to interpret self-potential data more accurately and e...The self-potential method is widely used in environmental and engineering geophysics. Four intelligent optimization algorithms are adopted to design the inversion to interpret self-potential data more accurately and efficiently: simulated annealing, genetic, particle swarm optimization, and ant colony optimization. Using both noise-free and noise-added synthetic data, it is demonstrated that all four intelligent algorithms can perform self-potential data inversion effectively. During the numerical experiments, the model distribution in search space, the relative errors of model parameters, and the elapsed time are recorded to evaluate the performance of the inversion. The results indicate that all the intelligent algorithms have good precision and tolerance to noise. Particle swarm optimization has the fastest convergence during iteration because of its good balanced searching capability between global and local minimisation.展开更多
基金the work is supported by scientific and technological fund of CNPC
文摘The problem of potential field inversion can be become that of solving system of linear equations by using of linear processing. There are a lot of algorithms for solving any system of linear equations, and the regularized method is one of the best algorithms. But there is a shortcoming in application with the regularized method, viz. the optimum regularized parameter must be determined by experience, so it is difficulty to obtain an optimum solution. In this paper, an iterative algorithm for solving any system of linear equations is discussed, and a sufficient and necessary condition of the algorithm convergence is presented and proved. The algorithm is convergent for any starting point, and the optimum solution can be obtained, in particular, there is no need to calculate the inverse matrix in the algorithm. The typical practical example shows the iterative algorithm is simple and practicable, and the inversion effect is better than that of regularized method.
文摘Cooperative path planning is an important area in fixed-wing UAV swarm.However,avoiding multiple timevarying obstacles and avoiding local optimum are two challenges for existing approaches in a dynamic environment.Firstly,a normalized artificial potential field optimization is proposed by reconstructing a novel function with anisotropy in each dimension,which can make the flight speed of a fixed UAV swarm independent of the repulsive/attractive gain coefficient and avoid trapping into local optimization and local oscillation.Then,taking into account minimum velocity and turning angular velocity of fixed-wing UAV swarm,a strategy of decomposing target vector to avoid moving obstacles and pop-up threats is proposed.Finally,several simulations are carried out to illustrate superiority and effectiveness.
基金Projects(30270496,60075019,60575012)supported by the National Natural Science Foundation of China
文摘A novel method was designed to solve reinforcement learning problems with artificial potential field.Firstly a reinforcement learning problem was transferred to a path planning problem by using artificial potential field(APF),which was a very appropriate method to model a reinforcement learning problem.Secondly,a new APF algorithm was proposed to overcome the local minimum problem in the potential field methods with a virtual water-flow concept.The performance of this new method was tested by a gridworld problem named as key and door maze.The experimental results show that within 45 trials,good and deterministic policies are found in almost all simulations.In comparison with WIERING's HQ-learning system which needs 20 000 trials for stable solution,the proposed new method can obtain optimal and stable policy far more quickly than HQ-learning.Therefore,the new method is simple and effective to give an optimal solution to the reinforcement learning problem.
基金Project(61273187)supported by the National Natural Science Foundation of ChinaProject(61321003)supported by the Foundation for Innovative Research Groups of the National Natural Science Foundation of China
文摘Region partition(RP) is the key technique to the finite element parallel computing(FEPC),and its performance has a decisive influence on the entire process of analysis and computation.The performance evaluation index of RP method for the three-dimensional finite element model(FEM) has been given.By taking the electric field of aluminum reduction cell(ARC) as the research object,the performance of two classical RP methods,which are Al-NASRA and NGUYEN partition(ANP) algorithm and the multi-level partition(MLP) method,has been analyzed and compared.The comparison results indicate a sound performance of ANP algorithm,but to large-scale models,the computing time of ANP algorithm increases notably.This is because the ANP algorithm determines only one node based on the minimum weight and just adds the elements connected to the node into the sub-region during each iteration.To obtain the satisfied speed and the precision,an improved dynamic self-adaptive ANP(DSA-ANP) algorithm has been proposed.With consideration of model scale,complexity and sub-RP stage,the improved algorithm adaptively determines the number of nodes and selects those nodes with small enough weight,and then dynamically adds these connected elements.The proposed algorithm has been applied to the finite element analysis(FEA) of the electric field simulation of ARC.Compared with the traditional ANP algorithm,the computational efficiency of the proposed algorithm has been shortened approximately from 260 s to 13 s.This proves the superiority of the improved algorithm on computing time performance.
基金Project(41574123)supported by the National Natural Science Foundation of ChinaProject(2015zzts250)supported by the Fundamental Research Funds for the Central Universities,ChinaProject(2013FY110800)supported by the National Basic Research Scientific Program of China
文摘The self-potential method is widely used in environmental and engineering geophysics. Four intelligent optimization algorithms are adopted to design the inversion to interpret self-potential data more accurately and efficiently: simulated annealing, genetic, particle swarm optimization, and ant colony optimization. Using both noise-free and noise-added synthetic data, it is demonstrated that all four intelligent algorithms can perform self-potential data inversion effectively. During the numerical experiments, the model distribution in search space, the relative errors of model parameters, and the elapsed time are recorded to evaluate the performance of the inversion. The results indicate that all the intelligent algorithms have good precision and tolerance to noise. Particle swarm optimization has the fastest convergence during iteration because of its good balanced searching capability between global and local minimisation.
