调用nlrx包simdesign_lhs时出现Atomic Vector Error求助

问题描述 #

在使用R的nlrx包为NetLogo模型集成优化功能时，执行nl@simdesign <- simdesign_lhs(nl=nl)语句时出现错误：

Error in x$min : $ operator is invalid for atomic vectors

相关代码如下：

library(nlrx)

root_d <- file.path("/home/imantha/workspace/CRTT")
netlogo_p <- file.path("/home/imantha/tools/NetLogo-6.4.0-64")
model_p <- file.path(root_d, "model","crtt_v2.nlogo")

# 创建结果存储目录
dir.create(file.path(root_d, "results"), showWarnings = FALSE)

out_p <- file.path(root_d, "results")

nl <- nl(nlversion = "6.4.0", nlpath = netlogo_p, modelpath = model_p, jvmmem = 1024)

# 绑定实验配置
nl@experiment <- experiment(
  expname = "crtt-test",
  outpath = out_p,
  repetition = 1,
  tickmetrics = "true",
  idsetup = "Setup",
  idgo = "Go",
  runtime = 7201,
  evalticks = seq(180,7200,60),
  metrics = c(
    "count cas-agents with [pma-name = 'Vieux Habitants PMA' and in-queue? = True and deceased? = False]"
  ),
  variables = list(
    "heli-rescue?" = FALSE,
    "Speed-Helicopter" = 200, 
    "no-heli-agent" = 3,
    "heli-agent-cap" = 2,
    
    "no-resc-agent-A" = 20,
    "no-resc-agent-B" = 15,
    "resc-agent-cap" = 3,
    "resc-agent-A-only" = TRUE,
    "Rescue-Time-Mins" = 10,
    "Transfer-Time-Mins" = 5,
    "Speed" = 30,
    "increase-rescue-agents-at-time" = 60,
    "Increase-Rescue-Agents" = 2,
    "transfer-hosp-vh" = 10,
    "transfer-hosp-sm" = 8,
    "pma-cap-t1" = 5,
    "pma-cap-t2" = 30,
    "pma-operational-t1" = 180,
    "pma-operational-t2" = 300,
    "stabilisation-t" = 17,
    "no-bb" = 6,
    "no-nbb" = 212
  ),
  constants = list()
)

nl@simdesign <- simdesign_lhs(nl=nl)

错误原因与解决方法 #

错误原因 #

simdesign_lhs()是生成拉丁超立方抽样设计的函数，要求**variables列表中的每个变量必须定义min和max取值范围**，这样函数才能生成有效抽样值。但当前代码里的variables所有元素都是单一固定值（如"no-heli-agent" = 3），属于原子向量，没有min属性，函数调用时尝试访问x$min就会触发错误。

解决步骤 #

1. 区分变量与常量：将不需要优化的固定值移到constants列表中，仅把需要抽样优化的变量留在variables里。
2. 为变量定义取值范围：将variables中的元素改为包含min和max的列表格式。

示例修改后的代码片段：

nl@experiment <- experiment(
  expname = "crtt-test",
  outpath = out_p,
  repetition = 1,
  tickmetrics = "true",
  idsetup = "Setup",
  idgo = "Go",
  runtime = 7201,
  evalticks = seq(180,7200,60),
  metrics = c(
    "count cas-agents with [pma-name = 'Vieux Habitants PMA' and in-queue? = True and deceased? = False]"
  ),
  # 仅保留需要优化的变量，并定义取值范围
  variables = list(
    "no-heli-agent" = list(min=1, max=5),
    "no-resc-agent-A" = list(min=10, max=30),
    "no-resc-agent-B" = list(min=10, max=20)
  ),
  # 固定值移到constants中
  constants = list(
    "heli-rescue?" = FALSE,
    "Speed-Helicopter" = 200, 
    "heli-agent-cap" = 2,
    "resc-agent-cap" = 3,
    "resc-agent-A-only" = TRUE,
    "Rescue-Time-Mins" = 10,
    "Transfer-Time-Mins" = 5,
    "Speed" = 30,
    "increase-rescue-agents-at-time" = 60,
    "Increase-Rescue-Agents" = 2,
    "transfer-hosp-vh" = 10,
    "transfer-hosp-sm" = 8,
    "pma-cap-t1" = 5,
    "pma-cap-t2" = 30,
    "pma-operational-t1" = 180,
    "pma-operational-t2" = 300,
    "stabilisation-t" = 17,
    "no-bb" = 6,
    "no-nbb" = 212
  )
)

# 现在可以正常生成抽样设计
nl@simdesign <- simdesign_lhs(nl=nl)

额外说明 #

• 对于布尔类型的变量（如heli-rescue?），如果需要优化，可定义min=FALSE, max=TRUE，函数会自动处理布尔值的抽样。
• 执行前确保NetLogo模型路径正确，且模型中的变量名与代码中的完全一致（注意大小写和特殊字符）。

内容的提问来源于stack exchange，提问作者imantha