Research And Application Of Improved Bat Algorithm In Optimization Of Reservoir Scheduling

Reservoir optimization scheduling is a high dimensional,multi stage nonlinear optimization problem.The factors that need to be comprehensively considered for such problems are also increasingly complex.With the increase of the complexity,the calculation of the optimal operation of the reservoir has also become more and more large,resulting in the “dimensional disaster”,and the traditional optimization method has become more difficult to solve.With many intelligent algorithms proposed and improved,they have been widely used in reservoir optimization scheduling problems.This paper proposes to use the bat algorithm and improve the constraint processing to solve such complex problems as reservoir optimization.This paper first discusses the significance and methods of mid-term and long term optimal operation of reservoirs and summarizes the research work at home and abroad.Then introduce the standard bat algorithm.Based on the standard bat algorithm(BA),this algorithm is used to calculate the poor convergence in the later period.In order to improve the performance of the algorithm and avoid premature phenomenon,the original algorithm is improved and an improved bat algorithm(IBA)is proposed.The improved bat optimization algorithm evenly scatters the initial population,increases the diversity of population initialization,and avoids falling into a local optimal solution during the optimization process.In order to improve the inertial coefficient,four updated forms of inertial coefficients were proposed,and a dynamic update method with better performance from large to small exponents was selected.The improved basic principles and calculation process were introduced.Several types of typical mathematics were introduced.A numerical simulation experiment was performed on the function,and the results were compared with the basic bat and related algorithms to illustrate the advantages of the improved bat optimization algorithm.Then,the improved bat algorithm is applied to the medium and long term optimal dispatch of single reservoirs,and the water level at the end of each month is optimized every year as a scheduling period.The optimization results(including the optimization of water level curves and power generation flow)are obtained.The basic bat algorithm and genetic algorithm are also used.Compared with the dynamic planning method,the benefits are obviously improved and the calculation speed is fast.In order to further verify the pros and cons of the algorithm,combined with the actual conditions for many years,the optimized conditions for the wet year,dry year,and flat year were given and compared.Finally,the improved bat algorithm was used to optimize two series of cascaded reservoirs,and the single-objective model with the largest total power generation and the multi-objective optimization model with the largest total power output for ensuring output were established.According to the bat algorithm optimization characteristics,the penalty function method combined with constraint processing is used to deal with equality constraints and inequality constraints,and the multi-objective is converted into a single object to obtain a joint fitness function.The concrete examples of cascade reservoirs show that,compared with the basic bat algorithm and related algorithms,the improved bat optimization algorithm has achieved good results and ensured an increase in output.The results show that the improved bat algorithm is suitable for solving such problems.Compared with the conventional intelligent algorithm and dynamic programming method,the optimization effect and calculation time have advantages,which provides a reference for the optimization of such problems.