CAJ | 학술논문

针对虚拟企业风险规划问题，在分析其各种风险具有随机性的特点的基础上，运用随机规划理论，分别建立风险规划的期望值模型和机会约束规划模型来描述决策者在不同风险偏好下的决策行为。针对所建立的模型，分别设计了基于蒙特卡罗模拟的粒子群优化算法、遗传算法和蚁群算法对其进行求解。仿真分析表明期望值模型较好地描述了风险中性决策者的决策行为，机会约束规划模型随着其偏好系数取值的不同描述了不同风险偏好（风险厌恶、风险中性、风险爱好）决策者的决策行为。通过对三种算法仿真结果的比较分析，表明基于蒙特卡罗模拟的粒子群优化算法在寻优能力、稳定性和收敛速度等方面优于其余两种算法，是解决此类风险规划问题的有效手段。
침대허의기업풍험규화문제，재분석기각충풍험구유수궤성적특점적기출상，운용수궤규화이론，분별건립풍험규화적기망치모형화궤회약속규화모형래묘술결책자재불동풍험편호하적결책행위。침대소건립적모형，분별설계료기우몽특잡라모의적입자군우화산법、유전산법화의군산법대기진행구해。방진분석표명기망치모형교호지묘술료풍험중성결책자적결책행위，궤회약속규화모형수착기편호계수취치적불동묘술료불동풍험편호（풍험염악、풍험중성、풍험애호）결책자적결책행위。통과대삼충산법방진결과적비교분석，표명기우몽특잡라모의적입자군우화산법재심우능력、은정성화수렴속도등방면우우기여량충산법，시해결차류풍험규화문제적유효수단。
For the stochastic characteristics of each risk for risk programming of virtual enterprise , with the sto-chastic programming theory , an expected value model and a chance constraint programming model are proposed to describe the decision behavior under various risk preferences .A Monte Carlo Simulation based Particle Swarm Optimization(MCS-PSO), Genetic Algorithm(MCS-GA)and Ant Colony Optimization(MCS-ACO)are designed to solve the models respectively .The simulation analysis shows that the expected value model describes the deci-sion behavior of risk-neutral decision maker , and the chance constraint programming model describes the decision behavior of decision makers with various risk preference (risk-neutral, risk-averse, risk-loving)while the prefer-ence coefficient has different values .The comparison of the simulation results from the three proposed algorithm shows that the MCS-PSO performs better than MCS-GA and MCS-ACO on searching ability , reliability and con-vergence speed , and MCS-PSO is an effective way to solve this kind of risk programming problems .