Research On Spectrum Resource Assignment Based On Hybrid Optimization Algorithm

With the rapid development of radio technology,the number and types of radio equipment continue to increase.How to use frequency resource assignment technology to improve spectrum utilization is a key issue in spectrum management.Especially in a modern information-based combat system with a complex electromagnetic environment,if there is a lack of effective and apposite spectrum management,different devices will interfere with each other at light,and the combat function will be collapsed at worst,and then the combat capability will be lost.In order to ensure the maximization of information-based combat effectiveness,it is necessary to ensure the normal operation of various radio equipment in the system.Therefore,it is particularly important to strengthen spectrum management and reasonably allocate spectrum resources within the system.The complex electromagnetic environment and the fast changing situation in modern war have put forward more efficient,dynamic and multi-dimensional requirements for spectrum management technology.Therefore,frequency assignment as a key technology of spectrum management has received extensive attention.Aiming at the problem that links are often treated equally in the frequency assignment model,an improved assignment model based on the MI-FAP(Minimum Interference Frequency Assignment Problem,MI-FAP)model is proposed in this thesis.First,the solution vector representing the assignment results of all links selects frequency points from the available frequency set of each link,making it possible to assign devices with different operating bands in a centralized manner.Secondly,the minimum frequency interval between the links is directly stored in the matrix representing the interference constraint relationship.If there is no interference relationship between the links,it is represented by 0,so that the constraint matrix does not need to be updated in real time with the change of the link frequency.Finally,according to the different importance of different devices and links in the system,an objective function of interference weighting of each link is proposed,which makes it easier for links with higher weights to be assigned non-interference frequencies.Aiming at the problems of low convergence accuracy and long time-consuming of the current frequency assignment algorithm,an improved frequency assignment algorithm based on Simulated Annealing is proposed in this thesis.First,the Simulated Annealing and Artificial Fish Swarm Algorithm currently used for frequency assignment is studied in this thesis,and their advantages and disadvantages is analyzed.Combined with the greedy mechanism of prey behavior in Artificial Fish Swarm Algorithm and the Metropolis criterion in Simulated Annealing,an improved algorithm that uses Metropolis criterion to control the screening times of greedy algorithm is proposed in this thesis.Finally,the comparison of the results of the static frequency assignment experiment simulation and the dynamic frequency assignment experiment simulation shows that the improved algorithm is better to the Simulated Annealing and the Artificial Fish Swarm Algorithm in terms of convergence accuracy and convergence speed.Aiming at the problem of two-dimensional resource assignment model,a mathematical model to describe the two-dimensional assignment of time slot-frequency is proposed in this thesis.First,the slot and frequency resources assigned to the link are constructed in the form of ordinal pairs in the solution vector.Then,under the premise that the interference constraint matrix is still consistent with the static frequency assignment problem,the objective function for calculating the interference cost is designed.According to in each slot whether different links occupy the slot,and the relationship of different links in the interference constraint matrix,the interference cost of a single time slot is calculated and then accumulated to achieve the purpose of calculating the total interference cost.Finally,through the experimental simulation results,it is proved that the improved algorithm proposed in this thesis can also be applied to the time slot-frequency two-dimensional assignment.