复杂网络提高的抗毁性优化设计仿真

摘要

实际的多智能体系统处在各种复杂的网络环境当中,复杂网络在面对随机失效时,常常表现出较强的抗毁性,而在面对恶意的攻击时,常常表现出较弱的抗毁性.如何提高和优化复杂网络的抗毁性是要面对和解决的问题,可以对复杂网络提高的抗毁性进行优化设计.具体可以用度分布熵和平均二步度来对复杂网络提高的抗毁性进行优化设计,分别通过遗传算法和非线性混合整数规划来解决约束条件下最优化度分布熵和无标度网络方面的问题.通过计算机仿真和理论推导,可以得到以下结论:度分布熵和平均二步度不仅可以作为网络抗毁性的量度而且还能帮助优化网络抗毁性.和度分布的熵相比,平均二步度不仅包含了度分布多样性的信息,而且还包含了网络拓扑的信息.因此,平均二步度是复杂网络异质性的一个更好的度量.%The actual multi-agent system in a variety of complex network environment,when faced random failure,often show a strong resistance to survivability;when faced malicious attacks,often show a weak resistance to survivability.How to improve and optimize the survivability of complex networks is the problem that we have to face and solve.We can optimize the design of improved survivability for complex network and can use degree distribution entropy and the average two-step degree to optimize the design of improved survivability for complex network.Genetic algorithm and nonlinear mixed integer programming are used to solve the problem of optimal degree distribution entropy and scale-free networks under the constraint condition.Through computer simulation and theoretical derivation,it can be found that degree distribution entropy and the average two-step degree not only can be used as a measure of the network survivability but also can help to optimize the network survivability.For degree distribution entropy,the average two-step degree as a complex network heterogeneity measure not only contains the diversity of degree distribution information,but also contains the information of network topology.Therefore,the average two-step degree is a better measures of the heterogeneity of complex networks.