For solving the problem of voltage fluctuation and the increase of network loss caused by the connection of small hydropower to the distribution network

the reactive power optimization mathematical model is set up with minimum network loss of distribution network as target function and voltage offset and reactive power compensation as penal function. The intelligent hybrid solution algorithm based on multi-agent and pseudo parallel genetic algorithm is proposed

which completes the update of all agents by way of exchange of adjacent set topology and grid environment. The parallelism of evolution is increased. The transmission data amount in the system is reduced in the system at the time of strengthening system diversity by way of introducing self-learning operators on the basis of optimizing genetic operator

thus improving the ability of global search of the genetic algorithm. Finally

the IEEE 30 node system is simulated and verified in Matlab and the obtained result is compared with the traditional genetic algorithm and multi-agent genetic algorithm. It is shown by the result that the algorithm can reduce effectively the network loss

improve voltage level and calculation speed

and have stronger global optimization and efficiency.