A Network Selection Scheme Based on Bee Colony Evolution.pptVIP

There are 2 bee colonies including one queen bee and many drones. Each individual IN=CH1,CH2,…,CHN is a solution of network selection problem in bee colonies. Chromosome CHn=ANn,Kn expresses access network ANn provides a Kn level service to the terminal. If in a solution or individual satisfy: then it is a feasible solution. Application type Coding Scheme Working frequency Signal strength Supporting velocity The fitness function as follow. Step 1: Initialization. Step 2: According to the feasible solution, generate one bee colony A, and make the terminal and access network play the game in every chromosome, then determine the utilities of both sides. Regard the individual with the smallest fitness value as the queen bee QA. Step 3: The same with Step 2, generate an other bee colony B, and get the queen bee QB. Step 4: Randomly select s1 and s2 individuals to execute the linear crossover operation with QA and QB from A and B in turn, where the sum of s1 and s2 is a half of colony scale. Then select 2×s1 and 2×s2 the best individuals from two bee colonies to form a colony C. Step 5: Bee colony C execute the mutation operation with ensure all new individuals are feasible, and the new bee colony is D. Step 6: Calculate the fitness value of individuals in D, and select the one with the smallest fitness value as the queen bee QD. Step 7: If the fitness value of QD is smaller than QA and QB, QD is the next generation queen bee, and D is the next generation bee colony A; Otherwise, replace the worst individual of D with the better one of QA and QB, then D is still the next generation bee colony A. Step 8: If the iteration time does not reach the maximum value, go to Step 3; otherwise, the individual with the smallest fitness value is the optimum solution, finish. Heterogeneous wireless network topologies User cases Simulate this proposed scheme (called Scheme 1), the greedy algorithm based scheme (called Scheme 2) and the QoS based scheme (called Scheme