基于逐次逼近遗传算法的梯级水库优化调度

利用传统遗传算法求解水库优化调度问题时,经过遗传操作产生的新个体可能是不可行解,因此需要对其进行修正.但在梯级水库调度中,由于各时段间、水库间存在的水力电力联系,使这种修正变得复杂困难.鉴于此,提出了逐次逼近遗传算法(GASA),它可在包含不可行解的空间中寻优,并通过搜索空间的不断改变,逐渐逼近最优解.最后通过一个算例,并与离散微分动态规划法(DDDP)和逐步优化法(POA)的优化结果进行比较,说明了该方法的可行性与有效性.

Cascade reservoirs optimal operation based on genetic algorithm successive approximation

When applying Genetic Algorithm(GA) to the problem of reservoir operation,the new chroms created by genetic operations are often infeasible and needed to revise.In cascade reservoirs operation,however,such revision becomes complicated because of the hydraulic and electric connections between time sequences and reservoirs.Therefore,an advanced GA—Genetic Algorithm Successive Approximation(GASA)—is proposed in this paper.Firstly,the possible optimum is obtained by traditional GA without modifying any infeasible alternatives,then a new research space is formed by adding or subtracting a value to the optimum,and new times of optimization are then carried on in altering new spaces until the conditions for procedure ending are satisfied.The advantage of this algorithm is that it can do optimizing within a space including feasible and infeasible schemes,which is attractive in the optimization of such complicated systems as cascaded hydropower plants.A simulated example is also provided,and the results are compared with those obtained by Discrete Differential Dynamic Programming(DDDP) and Progressive Optimality Algorithm(POA),which indicates the feasibility and validity of GASA.