Abstract: The mean square variance of resource is taken as an evaluation function in this paper to measure the balance of network planned resource Based on the resource intensity of non-critical work, which is used in different period of time, a strategy of dynamically parallel adjusting the flexible time of non-critical work and the associated resource intensity in each of the period in the same time is proposed, and an optimization model of network planned resource is also established Moreover, aiming at the feature of optimization model of network planning, which is a nonlinear optimal function with multi-peak values, global optimal solution is searched by using the modified differential evolution, which is based on the global searching of population, so that starting and ending times of every norr critical work can be optimized, and the optimal arrangement can be also obtained.At last, the superiority and the practicality of the new proposed optimization method are proved by doing case study and being compared with other algorithms.

Key words: resource leveling optimization, modified differential evolution(M D E), global optimal solution, resource intensity