NetLogo统计同类型相邻海龟组成的簇的大小问题求助

原有代码问题点 #

• 簇标记逻辑错误：逐个遍历绿海龟时只要发现block为0就生成新簇编号，没有按连通性合并同一簇的海龟，会导致同一个连通簇被分配多个编号
• 语法逻辑混乱：while条件中的嵌套调用主体不明确，myself指代错误，in-radius的调用主体和筛选对象关系错位，无法正确筛选到当前簇相邻的未标记绿海龟
• 执行时序问题：原代码中标记逻辑在ask turtles内部执行，海龟的标记操作会实时影响遍历逻辑，导致标记遗漏
• 原代码hatch逻辑缺少闭合方括号，运行前需要先补全语法漏洞

修复方案 #

首先删除rules过程中ask turtles块内的set block 0语句，将簇标记代码统一放在rules末尾、tick之前，代码如下：

; 每轮重新计算簇时先重置所有绿海龟的block值
ask turtles with [MHC = 1] [set block 0]
let current-block 1
; 循环处理所有未标记的绿海龟簇
while [any? turtles with [MHC = 1 and block = 0]] [
  ; 取当前未标记簇的起点
  let start one-of turtles with [MHC = 1 and block = 0]
  ask start [set block current-block]
  let current-cluster turtles with [block = current-block]
  ; 向外扩散标记同一簇的所有连通绿海龟
  while [any? turtles with [
    MHC = 1 and block = 0 and
    any? current-cluster in-radius 1
  ]] [
    ask turtles with [
      MHC = 1 and block = 0 and
      any? current-cluster in-radius 1
    ] [set block current-block]
    set current-cluster turtles with [block = current-block]
  ]
  set current-block current-block + 1
]

; 按需添加簇大小统计代码
let cluster-size-list map [i -> count turtles with [block = i]] n-values (current-block - 1) [? + 1]
print (word "当前绿海龟簇大小列表：" cluster-size-list)

优化说明 #

• 如果不需要每tick重新计算簇，可删除重置block的语句，仅在set-up阶段初始化block为0即可，大幅提升运行性能
• 如果是严格网格的四邻接（不包含对角线）判断，可将in-radius 1替换为any? current-cluster on neighbors，邻接判断更准确

内容的提问来源于stack exchange，提问作者James W