基于SVR-IPSO的高危行业企业安全投入优化模型及其改进——以煤炭生产企业为例

传统的安全投入模型对解决高危行业领域中模糊复杂的安全投入问题具有一定局限性，尤其当建立目标函数时，采用隐含线性关系假设的函数进行拟合会影响模型的推广能力。基于此，本文首先采用支持向量回归机（SVR）建立事故损失模型，与传统C-D函数拟合结果相比，该模型具有更好的预测能力；然后，以实际安全投入要求为约束，以安全总成本最小化为原则建立企业安全投入优化模型；最后，采用基于捕食搜索策略的粒子群算法对模型进行求解，同时，为保证全局收敛性，引入自适应控制策略对算法进行了改进。结果表明：该模型能够更加准确地描述安全投入与安全成本间的非线性作用关系，并通过粒子群寻优得到具备可行性的全局最优解，为高危行业企业安全投入结构优化提供新的决策思路。

Safety Investment Optimization Model of High Risk Industry Enterprises Based on SVR -IPSO and Its Improvement---A Case Study of Coal Production Enterprises

Input the traditional security model for solving the fuzzy and complex safety investment problem in high -risk industry has certain limitations . Especially when establishing the objective function , to use a function implied linear relation will affect the model’s gen-eralization ability . Based on this , this paper firstly used the support vector regression machine (SVR) to establish the accident loss mod-el , compared with the traditional C-D function fitting results , the model had a better predictive ability ;then , taking the actual security investment requirements as the constraint and the minimize safety cost as the principle , this paper established the enterprise security invest-ment optimization model ;finally , the model was solved by particle swarm optimization algorithm based on predatory search strategy . At the same time , in order to guarantee the global convergence , the algorithm was improved by introducing the adaptive control strategy . The re-sults showed that the model could more accurately describe the nonlinear relationship between safety input and safety cost , and the particle swarm optimization algorithm to get the global optimal solution . This study provided a new decision-making method to optimize the safety investment structure of high risk industries .