⭐️ 前言
小编之前写过一篇博文:求解三维装箱问题的启发式深度优先搜索算法(python),详述了基于空间选择的三维装箱算法。本文考虑了一个事实:在某些情况下,我们在摆放物品时,总是优先选择较低的平面,基于这个常识,本文提出一种基于平面选择的三维装箱算法。废话不多说,开始算法之旅吧。
⭐️ 块
块的定义及生成可以参考博文:求解三维装箱问题的启发式深度优先搜索算法(python),这里就不在赘述。
⭐️ 平面
“平面”指可用于摆放货物的面。初始平面就是箱的整个底面,放入第一批货物后,“平面”包括了同批货物顶面形成的面和箱底面空余的部分。
本文算法采用由底向上的方式完成物品的装载,既优先铺满底面,然后再向上堆放物品。大体过程是:首先由完全相同的货物组成“块”,然后用块自底向上依次填充所选择的目标平面,并重新生成若干新的平面,然后不断重复上述过程来完成最终的摆放方案。下图演示了一个用4个“块”进行填充的简单过程,每个块顶部都生成了一个的平面。在这个例子中,填充完毕后,从原先的1个平面,分成了8个新的平面。
⭐️ 选择平面
当所用容器仅有一个时,初始情况下只有一个备选平面,即箱底面。
选择装载目标平面时,按照以下几条准则依次进行判断:
1) 为避免堆积过高,影响货物堆放的稳定性,优先选择空间位置较低(即参考点z坐标较小)的平面; 2) 若几个平面参考点z坐标相同,则优先选择面积较小的平面,因为面积小的平面在后面可能更难使用; 3) 若这些平面的参考点z坐标与面积均相同,则优先选择相对较狭长的平面,同样因为狭长的平面在后面可能更加难以利用; 4) 若以上3点均相同,先比较它们参考点的x坐标,选择x坐标最小的一个。若x坐标相同,则选择y坐标最小的一个。
⭐️ 填充平面
为能够充分利用空间,物品块与目标平面组合的保留规则依次为:
1) 货物组成的“块”能最大限度地利用目标平面,即放入块后目标平面剩余面积最小; 2) 若剩余面积相同,比较块的体积,保留最大体积的块。
下文将该准则简称为准则。
⭐️ 合并平面
当所选择的平面不能容纳任何一个货物时,更有效地利用空间,需进行平面合并。一个平面只能和其相邻的平面合并,相邻平面指高度相同,即具有相同参考点z坐标,至少有一边平齐,如下图所示类似情况的平面(图中蓝色部分表示目标平面(gs),白色部分表示其相邻平面(ns)):
首先目标平面(gs)先后在平面列表和备用平面列表中依次顺序查找是否有相邻平面(ns)。平面列表指还未使用的新平面列表,备用平面列表指已经过计算不可能放入任何物品的平面的列表。当一个目标平面找到与其相邻的平面时,通过试合并决定是否保留这一合并,并进行正式合并。保留某次试合并的基本准则包括:
(1) 可以装入至少一个仍有剩余的物品(种类、方向不限); (2) 合并后新平面的面积是否比原先两个平面都大。
设试合并后新生产的两块平面分别为ms1和 ms2,判断是否保留该次试合并及正式合并的步骤为:
- 判断 ns 是否满足准则(1)。若不满足则执行第 2)步,若满足执行第 3)步;
- 判断 ms1 和 ms2 是否满足准则(1)。将满足的一个平面插入平面列表,不满足的放入备用平面列表。若均不满足准则(1),则判断 ms1 和 ms2 中是否有能够足准则(2)的,若有,则将ms1和ms2均放入备用平面列表;否则不保留此次试合并,重新开始寻找相邻平面;
- 计算在放入物品后gs和ns将浪费的总面积(ws1)。根据ms1和ms2满足准则(1)的情况,计算合并后可能浪费的面积之和(ws2)。设平面 xs 的面积为Sqr(xs),在该平面上放入物品后将浪费的面积为WsSqr(xs),则根据ms1和ms2的满足准则(1)的情况,ws2有以下几种情况:
若ws1>ws2,则执行平面合并程序,并更新平面和备用平面列表;若ws1<=ws2,则不保留此次试合并,继续寻找其他相邻平面; - 若对平面列表和备用平面列表搜索完毕侯,最终仍没有找到可合并的平面,则将目标平面gs从平面列表移入备用平面列表。
⭐️ 产生新平面
在放入一“块”后,原目标平面被分成3个子平面,一个位于块的顶部(下图中深色部分),另外两个可有下面两种方式(下图中Rsa1和Rsa2)。我们选择产生的子平面面积较大的一个方式,即比较两种方式中面积较大的两个子平面,选择有最大面积子平面的生成方式。以下图为例,若(a)Rsa1>Rsa2,(b)Rsb2>Rsb1,则比较 Rsa1与Rsb2,若Rsa1>Rsb2,选择(a)方式。产生的子平面放入平面列表中。
⭐️ 程序及运行结果(笔者python运行环境为python3.7)
import copy import sys from itertools import product import math from matplotlib import pyplot as plt from mpl_toolkits.mplot3d.art3d import Poly3DCollection import numpy as np MAX_GAP = 0 # 绘图相关函数 def plot_linear_cube(ax, x, y, z, dx, dy, dz, color='red', linestyle=None): xx = [x, x, x+dx, x+dx, x] yy = [y, y+dy, y+dy, y, y] kwargs = {"alpha": 1, "color": color, "linewidth":2.5, "zorder":2} if linestyle: kwargs["linestyle"] = linestyle ax.plot3D(xx, yy, [z]*5, **kwargs) ax.plot3D(xx, yy, [z+dz]*5, **kwargs) ax.plot3D([x, x], [y, y], [z, z+dz], **kwargs) ax.plot3D([x, x], [y+dy, y+dy], [z, z+dz], **kwargs) ax.plot3D([x+dx, x+dx], [y+dy, y+dy], [z, z+dz], **kwargs) ax.plot3D([x+dx, x+dx], [y, y], [z, z+dz], **kwargs) def cuboid_data(o, size=(1, 1, 1)): X = [[[0, 1, 0], [0, 0, 0], [1, 0, 0], [1, 1, 0]], [[0, 0, 0], [0, 0, 1], [1, 0, 1], [1, 0, 0]], [[1, 0, 1], [1, 0, 0], [1, 1, 0], [1, 1, 1]], [[0, 0, 1], [0, 0, 0], [0, 1, 0], [0, 1, 1]], [[0, 1, 0], [0, 1, 1], [1, 1, 1], [1, 1, 0]], [[0, 1, 1], [0, 0, 1], [1, 0, 1], [1, 1, 1]]] X = np.array(X).astype(float) for i in range(3): X[:, :, i] *= size[i] X += np.array(o) return X def plotCubeAt(positions, sizes=None, colors=None, **kwargs): if not isinstance(colors, (list, np.ndarray)): colors = ["C0"] * len(positions) if not isinstance(sizes, (list, np.ndarray)): sizes = [(1, 1, 1)] * len(positions) g = [] for p, s, c in zip(positions, sizes, colors): g.append(cuboid_data(p, size=s)) return Poly3DCollection(np.concatenate(g), facecolors=np.repeat(colors, 6), **kwargs) # 箱子类 class Box: def __init__(self, lx, ly, lz, weight=0.0, type=0): # 长 self.lx = lx # 宽 self.ly = ly # 高 self.lz = lz # 重 self.weight = weight # 类型 self.type = type def __str__(self): return "lx: {}, ly: {}, lz: {}, weight: {}, type: {}".format(self.lx, self.ly, self.lz, self.weight, self.type) # 块类 class Block: def __init__(self, lx, ly, lz, require_list=[], weight=0.0, box_rotate=False): # 长 self.lx = lx # 宽 self.ly = ly # 高 self.lz = lz # 需要的物品数量 self.require_list = require_list # 需要的物品重量 self.weight = weight # 体积 self.volume = 0 # 是否旋转 self.box_rotate = box_rotate def __str__(self): return "lx: %s, ly: %s, lz: %s, volume: %s, require: %s, weight: %s, box_rotate: %a" % (self.lx, self.ly, self.lz, self.volume, self.require_list, self.weight, self.box_rotate) def __eq__(self, other): return self.lx == other.lx and self.ly == other.ly and self.lz == other.lz and self.box_rotate == other.box_rotate and (np.array(self.require_list) == np.array(other.require_list)).all() # 平面类 class Plane: def __init__(self, x, y, z, lx, ly, height_limit=0): # 坐标 self.x = x self.y = y self.z = z # 长 self.lx = lx # 宽 self.ly = ly # 限高 self.height_limit = height_limit self.origin = None def __str__(self): return "x:{}, y:{}, z:{}, lx:{}, ly:{}, height_limit:{}".format(self.x, self.y, self.z, self.lx, self.ly, self.height_limit) def __eq__(self, other): return self.x == other.x and self.y == other.y and self.z == other.z and self.lx == other.lx and self.ly == other.ly # 判断是否与另一个平面相邻(z坐标相同,至少有一个边平齐),并返回合并后的两个平面 def adjacent_with(self, other): if self.z != other.z: return False, None, None # 矩形中心 my_center = (self.x + self.lx / 2, self.y + self.ly / 2) other_center = (other.x + other.lx / 2, other.y + other.ly / 2) # 矩形相邻时的中心距离 x_adjacent_measure = self.lx / 2 + other.lx / 2 y_adjacent_measure = self.ly / 2 + other.ly / 2 # 宽边相邻,长边对齐 if x_adjacent_measure + MAX_GAP >= math.fabs(my_center[0] - other_center[0]) >= x_adjacent_measure: if self.y == other.y and self.ly == other.ly: ms1 = Plane(min(self.x, other.x), self.y, self.z, self.lx + other.lx, self.ly) return True, ms1, None if self.y == other.y: ms1 = Plane(min(self.x, other.x), self.y, self.z, self.lx + other.lx, min(self.ly, other.ly)) if self.ly > other.ly: ms2 = Plane(self.x, self.y + other.ly, self.z, self.lx, self.ly - other.ly) else: ms2 = Plane(other.x, self.y + self.ly, self.z, other.lx, other.ly - self.ly) return True, ms1, ms2 if self.y + self.ly == other.y + other.ly: ms1 = Plane(min(self.x, other.x), max(self.y, other.y), self.z, self.lx + other.lx, min(self.ly, other.ly)) if self.ly > other.ly: ms2 = Plane(self.x, self.y, self.z, self.lx, self.ly - other.ly) else: ms2 = Plane(other.x, other.y, self.z, other.lx, other.ly - self.ly) return True, ms1, ms2 # 长边相邻,宽边对齐 if y_adjacent_measure + MAX_GAP >= math.fabs(my_center[1] - other_center[1]) >= y_adjacent_measure: if self.x == other.x and self.lx == other.lx: ms1 = Plane(self.x, min(self.y, other.y), self.z, self.lx, self.ly + other.ly) return True, ms1, None if self.x == other.x: ms1 = Plane(self.x, min(self.y, other.y), self.z, min(self.lx, other.lx), self.ly + other.ly) if self.lx > other.lx: ms2 = Plane(self.x + other.lx, self.y, self.z, self.lx - other.lx, self.ly) else: ms2 = Plane(self.x + self.lx, other.y, self.z, other.lx - self.lx, other.ly) return True, ms1, ms2 if self.x + self.lx == other.x + other.lx: ms1 = Plane(max(self.x, other.x), min(self.y, other.y), self.z, min(self.lx, other.lx), self.ly + other.ly) if self.lx > other.lx: ms2 = Plane(self.x, self.y, self.z, self.lx - other.lx, self.ly) else: ms2 = Plane(other.x, other.y, self.z, other.lx - self.lx, other.ly) return True, ms1, ms2 return False, None, None # 问题类 class Problem: def __init__(self, container: Plane, height_limit=sys.maxsize, weight_limit=sys.maxsize, box_list=[], num_list=[], rotate=False): # 初始最低水平面 self.container = container # 限高 self.height_limit = height_limit # 限重 self.weight_limit = weight_limit # 箱体列表 self.box_list = box_list # 箱体数量 self.num_list = num_list # 是否考虑板材旋转 self.rotate = rotate # 放置类 class Place: def __init__(self, plane: Plane, block: Block): self.plane = plane self.block = block def __eq__(self, other): return self.plane == other.plane and self.block == other.block # 装箱状态类 class PackingState: def __init__(self, plane_list=[], avail_list=[], weight=0.0): # 装箱计划 self.plan_list = [] # 可用箱体数量 self.avail_list = avail_list # 可用平面列表 self.plane_list = plane_list # 备用平面列表 self.spare_plane_list = [] # 当前排样重量 self.weight = weight # 当前排样体积 self.volume = 0 # 选择平面 def select_plane(ps: PackingState): # 选最低的平面 min_z = min([p.z for p in ps.plane_list]) temp_planes = [p for p in ps.plane_list if p.z == min_z] if len(temp_planes) == 1: return temp_planes[0] # 相同高度的平面有多个的话,选择面积最小的平面 min_area = min([p.lx * p.ly for p in temp_planes]) temp_planes = [p for p in temp_planes if p.lx * p.ly == min_area] if len(temp_planes) == 1: return temp_planes[0] # 较狭窄的 min_narrow = min([p.lx/p.ly if p.lx <= p.ly else p.ly/p.lx for p in temp_planes]) new_temp_planes = [] for p in temp_planes: narrow = p.lx/p.ly if p.lx <= p.ly else p.ly/p.lx if narrow == min_narrow: new_temp_planes.append(p) if len(new_temp_planes) == 1: return new_temp_planes[0] # x坐标较小 min_x = min([p.x for p in new_temp_planes]) new_temp_planes = [p for p in new_temp_planes if p.x == min_x] if len(new_temp_planes) == 1: return new_temp_planes[0] # y坐标较小 min_y = min([p.y for p in new_temp_planes]) new_temp_planes = [p for p in new_temp_planes if p.y == min_y] return new_temp_planes[0] # 将某平面从可用平面列表转移到备用平面列表 def disable_plane(ps: PackingState, plane: Plane): ps.plane_list.remove(plane) ps.spare_plane_list.append(plane) # 生成简单块 def gen_simple_block(init_plane: Plane, box_list, num_list, max_height, can_rotate=False): block_table = [] for box in box_list: for nx in np.arange(num_list[box.type]) + 1: for ny in np.arange(num_list[box.type] / nx) + 1: for nz in np.arange(num_list[box.type] / nx / ny) + 1: if box.lx * nx <= init_plane.lx and box.ly * ny <= init_plane.ly and box.lz * nz <= max_height - init_plane.z: # 该简单块需要的立体箱子数量 requires = np.full_like(num_list, 0) requires[box.type] = int(nx) * int(ny) * int(nz) # 简单块 block = Block(lx=box.lx * nx, ly=box.ly * ny, lz=box.lz * nz, require_list=requires) # 简单块填充体积 block.volume = box.lx * nx * box.ly * ny * box.lz * nz # 简单块重量 block.weight = int(nx) * int(ny) * int(nz) * box.weight block_table.append(block) if can_rotate: # 物品朝向选择90度进行堆叠 if box.ly * nx <= init_plane.lx and box.lx * ny <= init_plane.ly and box.lz * nz <= max_height - init_plane.z: requires = np.full_like(num_list, 0) requires[box.type] = int(nx) * int(ny) * int(nz) # 简单块 block = Block(lx=box.ly * nx, ly=box.lx * ny, lz=box.lz * nz, require_list=requires, box_rotate=True) # 简单块填充体积 block.volume = box.ly * nx * box.lx * ny * box.lz * nz # 简单块重量 block.weight = int(nx) * int(ny) * int(nz) * box.weight block_table.append(block) return block_table # 生成可行块列表 def gen_block_list(plane: Plane, avail, block_table, max_height, avail_weight=sys.maxsize): block_list = [] for block in block_table: # 块中需要的箱子数量必须小于最初的待装箱的箱子数量 # 块的尺寸必须小于放置空间尺寸 # 块的重量必须小于可放置重量 if (np.array(block.require_list) <= np.array(avail)).all() and block.lx <= plane.lx and block.ly <= plane.ly \ and block.lz <= max_height - plane.z and block.weight <= avail_weight: block_list.append(block) return block_list # 查找下一个可行块 def find_block(plane: Plane, block_list, ps: PackingState): # 平面的面积 plane_area = plane.lx * plane.ly # 放入块后,剩余的最小面积 min_residual_area = min([plane_area - b.lx * b.ly for b in block_list]) # 剩余面积相同,保留最大体积的块 candidate = [b for b in block_list if plane_area - b.lx * b.ly == min_residual_area] # 可用平面最大高度 max_plane_height = min([p.z for p in ps.plane_list]) _candidate = sorted(candidate, key=lambda x: x.volume, reverse=True) # if max_plane_height == 0: # # 第一次放置体积最大的块 # _candidate = sorted(candidate, key=lambda x: x.volume, reverse=True) # else: # # 选择平面时尽量使放置物品后与已经放置的物品平齐 # _candidate = sorted(candidate, key=lambda x: x.lz + plane.z - max_plane_height) # _candidate = sorted(candidate, key=lambda x: x.volume + ps.volume - 2440*1220*1000) return _candidate[0] # 裁切出新的剩余空间(有稳定性约束) def gen_new_plane(plane: Plane, block: Block): # 块顶部的新平面 rs_top = Plane(plane.x, plane.y, plane.z + block.lz, block.lx, block.ly) # 底部平面裁切 if block.lx == plane.lx and block.ly == plane.ly: return rs_top, None, None if block.lx == plane.lx: return rs_top, Plane(plane.x, plane.y + block.ly, plane.z, plane.lx, plane.ly - block.ly), None if block.ly == plane.ly: return rs_top, Plane(plane.x + block.lx, plane.y, plane.z, plane.lx - block.lx, block.ly), None # 比较两种方式中面积较大的两个子平面,选择有最大面积子平面的生成方式 rsa1 = Plane(plane.x, plane.y + block.ly, plane.z, plane.lx, plane.ly - block.ly) rsa2 = Plane(plane.x + block.lx, plane.y, plane.z, plane.lx - block.lx, block.ly) rsa_bigger = rsa1 if rsa1.lx * rsa1.ly >= rsa2.lx * rsa2.ly else rsa2 rsb1 = Plane(plane.x, plane.y + block.ly, plane.z, block.lx, plane.ly - block.ly) rsb2 = Plane(plane.x + block.lx, plane.y, plane.z, plane.lx - block.lx, plane.ly) rsb_bigger = rsb1 if rsb1.lx * rsb1.ly >= rsb2.lx * rsb2.ly else rsb2 if rsa_bigger.lx * rsa_bigger.ly >= rsb_bigger.lx * rsb_bigger.ly: return rs_top, rsa1, rsa2 else: return rs_top, rsb1, rsb2 # 计算平面浪费面积 def plane_waste(ps: PackingState, plane: Plane, block_table, max_height, max_weight=sys.maxsize): # 浪费面积 waste = 0 if plane: block_list = gen_block_list(plane, ps.avail_list, block_table, max_height, max_weight - ps.weight) if len(block_list) > 0: block = find_block(plane, block_list, ps) waste = plane.lx * plane.ly - block.lx * block.ly else: waste = plane.lx * plane.ly return waste # 判断平面是否可以放置物品 def can_place(ps: PackingState, plane: Plane, block_table, max_height, max_weight=sys.maxsize): if plane is None: return False block_list = gen_block_list(plane, ps.avail_list, block_table, max_height, max_weight - ps.weight) return True if len(block_list) > 0 else False # 用块填充平面 def fill_block(ps: PackingState, plane: Plane, block: Block): # 更新可用箱体数目 ps.avail_list = (np.array(ps.avail_list) - np.array(block.require_list)).tolist() # 更新放置计划 place = Place(plane, block) ps.plan_list.append(place) # 更新体积利用率 ps.volume = ps.volume + block.volume # 产生三个新的平面 rs_top, rs1, rs2 = gen_new_plane(plane, block) # 移除裁切前的平面 ps.plane_list.remove(plane) # 装入新的可用平面 if rs_top: ps.plane_list.append(rs_top) if rs1: ps.plane_list.append(rs1) if rs2: ps.plane_list.append(rs2) # 合并平面 def merge_plane(ps: PackingState, plane: Plane, block_table, max_height, max_weight=sys.maxsize): # print("合并平面开始了") for ns in ps.plane_list + ps.spare_plane_list: # 不和自己合并 if plane == ns: continue # 找相邻平面 is_adjacent, ms1, ms2 = plane.adjacent_with(ns) if is_adjacent: # print("有相邻平面呦") block_list = gen_block_list(ns, ps.avail_list, block_table, max_height, max_weight - ps.weight) # 相邻平面本身能放入至少一个剩余物品 if len(block_list) > 0: block = find_block(ns, block_list, ps) # 计算相邻平面和原平面浪费面积的总和 ws1 = ns.lx * ns.ly - block.lx * block.ly + plane.lx * plane.ly # 计算合并后平面的总浪费面积 ws2 = plane_waste(ps, ms1, block_table, max_height, max_weight) + plane_waste(ps, ms2, block_table, max_height, max_weight) # 合并后,浪费更小,则保留合并 if ws1 > ws2: # 保留平面合并 ps.plane_list.remove(plane) if ns in ps.plane_list: ps.plane_list.remove(ns) else: ps.spare_plane_list.remove(ns) if ms1: ps.plane_list.append(ms1) if ms2: ps.plane_list.append(ms2) return else: # 放弃平面合并,寻找其他相邻平面 continue # 相邻平面本身无法放入剩余物品 else: # 合共后产生一个平面 if ms2 is None: # 能放物品,则保留平面合并 if can_place(ps, ms1, block_table, max_height, max_weight): ps.plane_list.remove(plane) if ns in ps.plane_list: ps.plane_list.remove(ns) else: ps.spare_plane_list.remove(ns) ps.plane_list.append(ms1) return elif ms1.lx * ms1.ly > plane.lx * plane.ly and ms1.lx * ms1.ly > ns.lx * ns.ly: ps.plane_list.remove(plane) if ns in ps.plane_list: ps.plane_list.remove(ns) else: ps.spare_plane_list.remove(ns) # ps.spare_plane_list.append(ms1) ps.plane_list.append(ms1) return else: continue # 合并后产生两个平面 else: if (not can_place(ps, ms1, block_table, max_height, max_weight)) and (not can_place(ps, ms2, block_table, max_height, max_weight)): if (ms1.lx * ms1.ly > plane.lx * plane.ly and ms1.lx * ms1.ly > ns.lx * ns.ly) or (ms2.lx * ms2.ly > plane.lx * plane.ly and ms2.lx * ms2.ly > ns.lx * ns.ly): ps.plane_list.remove(plane) if ns in ps.plane_list: ps.plane_list.remove(ns) else: ps.spare_plane_list.remove(ns) ps.spare_plane_list.append(ms1) ps.spare_plane_list.append(ms2) return else: continue else: ps.plane_list.remove(plane) if ns in ps.plane_list: ps.plane_list.remove(ns) else: ps.spare_plane_list.remove(ns) if can_place(ps, ms1, block_table, max_height, max_weight): ps.plane_list.append(ms1) else: ps.spare_plane_list.append(ms1) if can_place(ps, ms2, block_table, max_height, max_weight): ps.plane_list.append(ms2) else: ps.spare_plane_list.append(ms2) return # 若对平面列表和备用平面列表搜索完毕后,最终仍没有找到可合并的平面,则将目标平面从平面列表移入备用平面列表 disable_plane(ps, plane) # 构建箱体坐标,用于绘图 def build_box_position(block, init_pos, box_list): # 箱体类型索引 box_idx = (np.array(block.require_list) > 0).tolist().index(True) if box_idx > -1: # 所需箱体 box = box_list[box_idx] # 箱体的相对坐标 if block.box_rotate: nx = block.lx / box.ly ny = block.ly / box.lx x_list = (np.arange(0, nx) * box.ly).tolist() y_list = (np.arange(0, ny) * box.lx).tolist() else: nx = block.lx / box.lx ny = block.ly / box.ly x_list = (np.arange(0, nx) * box.lx).tolist() y_list = (np.arange(0, ny) * box.ly).tolist() nz = block.lz / box.lz z_list = (np.arange(0, nz) * box.lz).tolist() # 箱体的绝对坐标 dimensions = (np.array([x for x in product(x_list, y_list, z_list)]) + np.array([init_pos[0], init_pos[1], init_pos[2]])).tolist() # 箱体的坐标及尺寸 if block.box_rotate: return sorted([d + [box.ly, box.lx, box.lz] for d in dimensions], key=lambda x: (x[0], x[1], x[2])), box_idx else: return sorted([d + [box.lx, box.ly, box.lz] for d in dimensions], key=lambda x: (x[0], x[1], x[2])), box_idx return None, None # 绘制排样结果 def draw_packing_result(problem: Problem, ps: PackingState): # 绘制结果 fig = plt.figure() ax1 = fig.gca(projection='3d') # 绘制容器 plot_linear_cube(ax1, 0, 0, 0, problem.container.lx, problem.container.ly, problem.height_limit) for p in ps.plan_list: # 绘制箱子 box_pos, _ = build_box_position(p.block, (p.plane.x, p.plane.y, p.plane.z), problem.box_list) positions = [] sizes = [] colors = ["yellow"] * len(box_pos) for bp in sorted(box_pos, key=lambda x: (x[0], x[1], x[2])): positions.append((bp[0], bp[1], bp[2])) sizes.append((bp[3], bp[4], bp[5])) pc = plotCubeAt(positions, sizes, colors=colors, edgecolor="k") ax1.add_collection3d(pc) plt.title('Cube{}'.format(0)) plt.show() # plt.savefig('3d_lowest_plane_packing.png', dpi=800) # 基本启发式算法 def basic_heuristic(problem: Problem): # 生成简单块 block_table = gen_simple_block(problem.container, problem.box_list, problem.num_list, problem.height_limit, problem.rotate) # 初始化排样状态 ps = PackingState(avail_list=problem.num_list) # 开始时,剩余空间堆栈中只有容器本身 ps.plane_list.append(Plane(problem.container.x, problem.container.y, problem.container.z, problem.container.lx,problem.container.ly)) max_used_high = 0 # 循环直到所有平面使用完毕 while ps.plane_list: # 选择平面 plane = select_plane(ps) # 查找可用块 block_list = gen_block_list(plane, ps.avail_list, block_table, problem.height_limit, problem.weight_limit - ps.weight) if block_list: # 查找下一个近似最优块 block = find_block(plane, block_list, ps) # 填充平面 fill_block(ps, plane, block) # 更新排样重量 ps.weight += block.weight # 更新最大使用高度 if plane.z + block.lz > max_used_high: max_used_high = plane.z + block.lz else: # 合并相邻平面 merge_plane(ps, plane, block_table, problem.height_limit, problem.weight_limit) # 板材的位置信息 box_pos_info = [[] for _ in problem.num_list] for p in ps.plan_list: box_pos, box_idx = build_box_position(p.block, (p.plane.x, p.plane.y, p.plane.z), problem.box_list) for bp in box_pos: box_pos_info[box_idx].append((bp[0], bp[1], bp[2])) # 计算容器利用率 used_volume = problem.container.lx * problem.container.ly * max_used_high used_ratio = round(float(ps.volume) * 100 / float(used_volume), 3) if used_volume > 0 else 0 # # 绘制排样结果图 draw_packing_result(problem, ps) return ps.avail_list, used_ratio, max_used_high, box_pos_info, ps # 主算法 def simple_test(): # 容器底面 container = Plane(0, 0, 0, 2440, 1220) # 箱体列表 box_list = [Box(lx=2390, ly=70, lz=10, weight=3001, type=0), Box(lx=2390, ly=50, lz=10, weight=10, type=1), Box(lx=625, ly=210, lz=10, weight=5, type=2), Box(lx=625, ly=110, lz=10, weight=5, type=3), Box(lx=2160, ly=860, lz=10, weight=5, type=4), Box(lx=860, ly=140, lz=10, weight=5, type=5), Box(lx=860, ly=120, lz=10, weight=5, type=6)] num_list = [200, 3, 200, 180, 20, 5, 10] # 问题 problem = Problem(container=container, height_limit=300, weight_limit=4000, box_list=box_list, num_list=copy.copy(num_list)) # 具体计算 new_avail_list, used_ratio, used_high, box_pos_, _ = basic_heuristic(problem) # 箱体原始数量 print(num_list) # 剩余箱体 print(new_avail_list) # 利用率 print(used_ratio) if __name__ == "__main__": simple_test()
算法运行结果如下:
⭐️ 写在最后
本文理论部分参考上海交通大学硕士论文《三维装箱问题的混合遗传算法研究》,论文作者和导师都很牛,读者可以自行百度获取论文原文😏。
笔者水平有限,若有不对的地方欢迎评论指正!