违规放牧行为的多智能体系统模拟

The simulation of illegal grazing with multi-agent systems

退牧还草政策实施过程中出现牧户违规放牧行为。违规放牧是在禁牧或休牧期进行放牧。本研究构建牧户-政府行为的多智能体模型,模拟不同政策措施情景下的牧户放牧行为及其预期收益。研究结果表明,其他因素不变情况下,检查人数与牧户数之比达到1.5,或者违规放牧被检查到的概率大于或等于0.73,或者罚款金额和牧户违规放牧收益的比值达到5.8,均可使得牧户收益为零。这意味着在一定范围内增加检查执法人数、提高检查频率、提高罚款金额有助于减少违规放牧行为。本研究仅在降低牧户经济收益的框架下,讨论相应政策措施的效果。本研究中提出的建议可在一定程度上减少违规放牧行为,要想从根本上杜绝违规放牧行为,需要转变牧民生产方式,提高牧民生活水平与质量。本研究结果可为制定政策与措施以减少或缓解违规放牧行为提供依据与参考。

Illegal grazing,i.e.,the grazing in the prohibiting or resting periods,occurred in the implementation of the policy‘returning grazing land to grassland’.In this study,a herdsman-government multi-agent model was build to simulate several scenarios of policy measures with adjusting the parameters.The results showed that the illegal grazing could be effectively controlled in the conditions that the ratio of the number of inspectors and the number of herdsmen reaches 1.5 or the probability of inspecting equals to or more than0.73 or the ratio of the penalty amount and corresponding benefit reaches 5.8.It is indicated that,to reduce illegal grazing,increasing number of inspectors,and increasing inspecting frequency,and raising penalty amount are useful measures.The effect of policy measures discussed in this study will work in the framework that herdsmen regulate their behavior according to the quantity of economy benefits.The suggestions proposed in this study could not completely reduce the illegal grazing behavior.In order to fundamentally eliminate illegal grazing behavior,it is necessary to change the herdsmen’s production mode and improve the living standards and quality of herdsmen.This study is helpful for regulating policies and measures to reduce illegal grazing.