一种用于机组组合问题的改进双重粒子群算法

为了更经济快速地解决机组组合问题,提出一种改进双重粒子群优化(particle swarm optimization,PSO)算法,包含离散部分和连续部分。离散PSO分时段优化机组的启停状态,在种群更新时加入了临界算子,改进了可行解的判别条件,各机组出力最低值的和要在一定程度上低于负荷需求值,并考虑机组启停时间的向前继承和向后约束。连续PSO用于启停状态确定过程中和确定后的负荷分配,考虑功率平衡约束、热备用约束和机组的出力上下限约束。求解经济负荷分配时,利用罚函数的方法满足机组的爬坡速率约束,最后得到煤耗最小值。采用2个24时段的算例进行仿真,实验结果表明新算法减少了搜索量,提高了收敛速度,并为机组组合问题提出了新思路。

An Improved Dual Particle Swarm Optimization Algorithm for Unit Commitment Problem

To solve the unit commitment problem economically and quickly,an improved dual particle swarm optimization(PSO) algorithm including both discrete and continuous parts was proposed.The starting and shutdown state of units were optimized according to different period of time using discrete PSO,and a pair of critical operators was added into the algorithm;in addition,the criterion condition of feasible solution was modified,where the sum of each unit’s lowest value must be smaller than the load to some extent.The inheritance from earlier state and constraints to later period of time for running time and shutdown time were considered.The continuous PSO was used in units’ load dispatch during the process of deciding starting-stopping states and after the solution,where constraints of power balance,spinning reserve and lower and upper limits were considered.While solving the economic load dispatch,penalty function was adopted to satisfy the ramp rate constraints,and the minimum coal consumptions could be gained.Two examples including 24 period of time were simulated,the experimental results of which showed the proposed approach decreased amount of effort during search and improved convergence rate.In addition,the new method suggests new thinking for unit commitment problem.