如何高效实现效用值0-1区间截断？参数敏感性分析优化

我需要为一组参数分别计算基准值上下20%的数值，示例代码如下：

d_e
Minimum_d_e <- d_e - 0.20*d_e
Maximum_d_e <- d_e + 0.20*d_e

得到各参数的最小、最大值后，我创建了两个包含所有参数极值的向量：

min_vec  = c(Minimum_HR_FP_Exp, Minimum_HR_FP_SoC, Minimum_HR_PD_SoC, Minimum_HR_PD_Exp, 
             Minimum_P_OSD_SoC, Minimum_P_OSD_Exp, Minimum_p_FA1_STD, Minimum_p_FA2_STD, 
             Minimum_p_FA3_STD, Minimum_p_FA1_EXPR, Minimum_p_FA2_EXPR, Minimum_p_FA3_EXPR, 
             Minimum_administration_cost, Minimum_c_PFS_Folfox, Minimum_c_PFS_Bevacizumab, 
             Minimum_c_OS_Folfiri, Minimum_c_AE1, Minimum_c_AE2, Minimum_c_AE3, 
             Minimum_d_e, Minimum_d_c, Minimum_u_F, Minimum_u_P, 
             Minimum_AE1_DisUtil, Minimum_AE2_DisUtil, Minimum_AE3_DisUtil)

max_vec  = c(Maximum_HR_FP_Exp, Maximum_HR_FP_SoC, Maximum_HR_PD_SoC, Maximum_HR_PD_Exp, 
             Maximum_P_OSD_SoC, Maximum_P_OSD_Exp, Maximum_p_FA1_STD, Maximum_p_FA2_STD, 
             Maximum_p_FA3_STD, Maximum_p_FA1_EXPR, Maximum_p_FA2_EXPR, Maximum_p_FA3_EXPR, 
             Maximum_administration_cost, Maximum_c_PFS_Folfox, Maximum_c_PFS_Bevacizumab, 
             Maximum_c_OS_Folfiri, Maximum_c_AE1,  Maximum_c_AE2, Maximum_c_AE3, 
             Maximum_d_e, Maximum_d_c, Maximum_u_F, Maximum_u_P, 
             Maximum_AE1_DisUtil, Maximum_AE2_DisUtil, Maximum_AE3_DisUtil)         

其中像Maximum_AE3_DisUtil这类效用值需要限制在0到1之间。我原本用手动替换的方式处理：

Maximum_AE3_DisUtil<- replace(Maximum_AE3_DisUtil, Maximum_AE3_DisUtil<0, 0)
Maximum_AE3_DisUtil<- replace(Maximum_AE3_DisUtil, Maximum_AE3_DisUtil>1, 1)

这种方法虽然可行，但效率很低，想找更高效的实现方式。

方法1：用pmin()和pmax()直接处理 #

这是R里最简洁高效的方式，通过嵌套函数直接把数值限制在[0,1]区间：

# 单个效用值处理
Maximum_AE1_DisUtil <- pmax(pmin(Maximum_AE1_DisUtil, 1), 0)
Maximum_AE2_DisUtil <- pmax(pmin(Maximum_AE2_DisUtil, 1), 0)
Maximum_AE3_DisUtil <- pmax(pmin(Maximum_AE3_DisUtil, 1), 0)

如果要批量处理多个效用值，可以打包成向量一次性操作：

util_values <- c(Maximum_AE1_DisUtil, Maximum_AE2_DisUtil, Maximum_AE3_DisUtil)
util_values_clamped <- pmax(pmin(util_values, 1), 0)
# 把处理结果赋值回原变量
c(Maximum_AE1_DisUtil, Maximum_AE2_DisUtil, Maximum_AE3_DisUtil) <- util_values_clamped

方法2：封装自定义限制函数 #

如果需要多次复用这个逻辑，可以写一个自定义函数：

clamp <- function(x, lower = 0, upper = 1) {
  pmax(pmin(x, upper), lower)
}

# 单个变量调用
Maximum_AE1_DisUtil <- clamp(Maximum_AE1_DisUtil)
Maximum_AE2_DisUtil <- clamp(Maximum_AE2_DisUtil)
Maximum_AE3_DisUtil <- clamp(Maximum_AE3_DisUtil)

# 批量处理多个变量
util_list <- list(Maximum_AE1_DisUtil, Maximum_AE2_DisUtil, Maximum_AE3_DisUtil)
util_list_clamped <- lapply(util_list, clamp)
# 赋值回原变量
c(Maximum_AE1_DisUtil, Maximum_AE2_DisUtil, Maximum_AE3_DisUtil) <- util_list_clamped

方法3：创建极值向量时直接处理 #

如果想在生成max_vec的阶段就完成范围限制，可以直接在向量定义中插入函数调用：

max_vec  = c(Maximum_HR_FP_Exp, Maximum_HR_FP_SoC, Maximum_HR_PD_SoC, Maximum_HR_PD_Exp, 
             Maximum_P_OSD_SoC, Maximum_P_OSD_Exp, Maximum_p_FA1_STD, Maximum_p_FA2_STD, 
             Maximum_p_FA3_STD, Maximum_p_FA1_EXPR, Maximum_p_FA2_EXPR, Maximum_p_FA3_EXPR, 
             Maximum_administration_cost, Maximum_c_PFS_Folfox, Maximum_c_PFS_Bevacizumab, 
             Maximum_c_OS_Folfiri, Maximum_c_AE1,  Maximum_c_AE2, Maximum_c_AE3, 
             Maximum_d_e, Maximum_d_c, Maximum_u_F, Maximum_u_P, 
             clamp(Maximum_AE1_DisUtil), clamp(Maximum_AE2_DisUtil), clamp(Maximum_AE3_DisUtil))         

内容的提问来源于stack exchange，提问作者James Moore