Comprehensive measures for the estimation performance evaluation (EPE) has become increasingly prominent. This paper proposed a new radar chart evaluation method to measure the estimation performance. Firstly, the new radar chart index, which is composed of several popular incomprehensive measures, are presented, and the method of the weight of the each index is calculated based on vector ranking method. Secondly, the new comprehensive measures for the EPE is designed according to the fan area and the fan arc length. Finally, two cases study are provided to verify the effectiveness of this method.
21 October 2022 in IEEE Transactions on Cloud Computing
To overcome the drawback of optimizing process with biased, slower convergence speed and easily falling into local optimum on high-dimensional optimization problems of Sparrow Search Algorithm (SSA), an improved Sparrow Search Algorithm (CRSSA) based on the Normal Cloud Model (NCM) and niche recombination strategy is proposed. Firstly, the position update strategy of the producers with alarm value less than safe value based on the NCM is proposed for avoiding the original update strategy to converge to the center of domain gradually; Secondly, the adaptive parameter $En$ in the NCM can better balance the exploitation and exploration capabilities of the optimization process; Then, the position update strategy of the followers with poor fitness value based on the NCM is presented for enhancing the diversity of population at the end of iteration and avoiding the algorithm to fall into local optimum; Finally, when the algorithm is stagnating, a niche-based recombination strategy is used to further avoid falling into the local optimum, and accelerating the convergence speed. Simulation results show that the CRSSA can not only avoid optimizing with biased of SSA, but keep better accuracy and convergence speed for solving high-dimensional complex optimization problems.
Lighting system scheme selection is an important part of infrastructure construction. This paper proposes an improved radar chart evaluation method for lighting system performance evaluation. First, to solve the problem of other weight scale methods’ inaccurate calculation, Pitman’s closeness measure in exponential form is used to determine the weight of the lighting system performance evaluation index. Second, a comprehensive performance evaluation model, applying the average area, average girth, and average eccentricity of the radar chart, is presented to solve the problem of inconsistent evaluation results due to the different orders of indicators. Finally, an example is given to verify the correctness and rationality of the proposed model. The research work provides scientific guidance for lighting system scheme selection.