Measurement, evaluation and algorithm selection on network planning problem

To investigate the treatment effect of different algorithms on network planning problem, the coefficients of network complexity (CNC) and network evaluation indices are proposed, taking the topological structure, interchangeability of resource and due date of the network into consideration. And then the relationship between the effects of different algorithms and the characteristics of network is analyzed based on multiple regression technology. The experimental results show that the proposed coefficients of network complexity are reasonable and the regression equations can accurately estimate the results of the algorithms, so that useful information can be provided to select a proper algorithm based on the characteristics of network planning problem.