CAJ | 학술논문

资源约束项目调度问题是工程管理领域研究的热点之一，但无论是模型构建还是求解均有一定的难度，尤其是模型求解已被证明是NP-hard问题。鉴于此，构建了以工期最短为优化目标的项目调度模型，为便于求解，将模型的显性约束和隐性约束做了适当处理，并利用差异演化算法较强的记忆能力和全局收敛能力以及模拟退火的局部跳出能力，将模拟退火算法和差异演化算法进行有效结合。通过工程实例，分别采用遗传算法、差异演化算法以及模拟退火差异演化算法进行求解。结果表明，3种算法都可以收敛到最优解，但论文算法具有较大的搜素范围与局部寻优能力，同时求解的稳定性指标明显优于遗传算法和差异演化算法。
자원약속항목조도문제시공정관리영역연구적열점지일，단무론시모형구건환시구해균유일정적난도，우기시모형구해이피증명시NP-hard문제。감우차，구건료이공기최단위우화목표적항목조도모형，위편우구해，장모형적현성약속화은성약속주료괄당처리，병이용차이연화산법교강적기억능력화전국수렴능력이급모의퇴화적국부도출능력，장모의퇴화산법화차이연화산법진행유효결합。통과공정실례，분별채용유전산법、차이연화산법이급모의퇴화차이연화산법진행구해。결과표명，3충산법도가이수렴도최우해，단논문산법구유교대적수소범위여국부심우능력，동시구해적은정성지표명현우우유전산법화차이연화산법。
Resource constrained project scheduling problem is one of the hotspots in research of project management. But both models building and solving have the certain difficulty. Especially models solving has been proved to be NP-hard problem. In view of this,this article constructed the project scheduling model of the shortest time limit as the optimization goal after explicit and implicit constraints of models have been appropriately treated for solving. Simulated annealing algorithm and differential evolution is combined effectively to take advantage of strong ability of memory and global convergence ability of differential evolution and the local jumping out of the ability of and simulated annealing. Through engineering examples,genetic algorithm,difference evolution algorithm and simulated annealing difference evolution algorithm have been programmed to solve respectively. Results show that the three kinds of algorithm can converge to the optimal solution,but the paper algorithm has a larger scope of search and local optimization ability,and solving stability index is superior to genetic algorithm and differential evolutionary algorithm.