[0026] ＜第一實施形態＞ 以下，參照圖1~圖13來說明本發明的一實施形態的配送計畫系統。 圖1是表示本發明的第一實施形態的配送計畫系統的一例的機能方塊圖。在本實施形態中，配送計畫系統是例如藉由1台的PC或伺服器裝置等的電腦裝置來構成。電腦裝置是包含CPU(Central Processing Unit)等的運算部及ROM(Read Only Memory)、RAM(Random Access Memory)、HDD(Hard Disk Drive)等的記憶部以及網路介面等其他的硬體而構成。 [0027] 圖1的配送計畫裝置10是配送計畫系統的一例。配送計畫裝置10是對於包含搭乘輸送的配送計畫，算出使成本形成最小的配送手段、配送路徑等的裝置。在本實施形態中，舉在下車離開型的汽車共享中將使用者共同利用的車輛配送往使用者開始利用的場所的場面，擬定實現該配送的最適的配送計畫之方法為例。配送配送物的情況，例如，被要求選擇使成本形成最小的配送手段、配送路徑。有關使成本形成最小的配送計畫的擬定是至今有各式各樣的方法被提供。但，在汽車共享的車輛的配送，例如，有與快遞等配送貨物的情況不同的點。那是在配送車輛時，人可搭乘於車輛移動的點。例如，在據點A、據點B、據點C、據點D、據點E的各個的據點，有車輛多餘的狀態或不足的狀態。如此的狀態中，為了從車輛多餘的據點往不足的據點移動車輛，因應使用者的需求，例如可思考以下般的方法。(1)一人的配送人員以可裝載車輛的卡車巡迴各據點，將多餘的車輛搭載於卡車，配送至有不足的據點。(2)複數的配送人員搭乘於配送車，移動於各據點。搭乘於配送車的配送人員的一部分是一旦到達車輛多餘的據點，則換乘至剩餘的車輛，將該車輛駕駛配送至車輛不足的據點(搭乘輸送)。如此的情況，不容易得知到底以哪個配送手段來進行配送為佳，以怎樣的路徑來進行配送，可使成本形成最小。本實施形態的配送計畫裝置10是對於可搭乘輸送時的配送計畫，導入根據數學見解的數理模型或限制，藉此提供一種高速且有效率地產生例如使成本形成最小的配送計畫之方法。本實施形態的配送計畫裝置10是在產生將配送物配送至配送據點的配送計畫時，根據作為該配送的初期條件被賦予之每個配送據點的配送物的需要數量(不足的數量)及供給數量(多餘的數量)來將初期條件所示的配送問題的全體分割成部分性的規模小的配送問題。藉此，即使配送的車輛或據點的數量為大規模，也可以實用性的時間(例如10分鐘)擬定配送計畫。 [0028] 如圖1所示，配送計畫裝置10是具備：初期條件設定部11、配送計畫產生部12、輸出入部13、第一區域分割部14及記憶部15。 初期條件設定部11是以配送主體(例如配送人員)可搭乘移動的配送物(例如使用者所利用的車輛)、配送主體、可移動配送物或配送主體的配送手段(例如卡車等的配送車)的任一個所停留的場所的每個配送據點的配送物的需要數量及供給數量，及表示配送主體與配送手段的初期位置的一個或複數的出發據點(例如提供汽車共享的服務的企業的運用據點)的資訊，及出發據點的可利用的配送手段和配送主體的資訊，以及配送期限的資訊等的資訊，作為在產生配送計畫上的初期條件設定。有關該等的參數是在之後詳細說明。 配送計畫產生部12是計算點資訊及分支資訊，至少產生1個將在配送期限內符合需要數量的配送物配送至設定有該需要數量的配送據點的情況的分支資訊的集合，該點資訊是將配送據點及出發據點與以配送開始時為基準的時刻設為一組，該分支資訊是點資訊之中涉及配送的2個點資訊之間的表示涉及配送的配送物及配送主體以及配送手段的流量。將出發據點及配送據點總稱為據點。 輸出入部13是受理使用者的輸入操作。輸出入部13是將根據配送計畫產生部12所產生的分支資訊的集合之配送計畫的資訊等輸出至顯示器等。 第一區域分割部14是使有需要配送物的配送據點群所含的一個的需要點，及成為配送物的供給源頭的配送據點群所含的一個的供給點，以從該一個的需要點往該一個的供給點的移動時間會成為最小的方式附上對應，產生附上對應的需要點及供給點的組合的集合。亦即，第一區域分割部14是將朝根據初期條件的有需要配送物的全部的配送據點群之配送問題的全體分割成產生的集合所示的配送區域單位的配送問題(從集合內所含的需要點群往供給點群的配送問題)。然後，配送計畫產生部12是針對分割後的規模小的配送問題的各者產生分支資訊的集合。 記憶部15是記憶配送計畫的產生所必要的諸資訊。 初期條件設定部11、配送計畫產生部12、第一區域分割部14是例如藉由配送計畫裝置10所具備的CPU(Central Processing Unit；中央處理裝置)從記憶部15讀出程式實行而實現。 [0029] 配送計畫裝置10是對於給予的配送問題，利用時空網路模型來產生配送計畫。首先，說明有關時空網路模型及利用時空網路模型的配送計畫的產生方法，然後，說明有關問題形成大規模時的配送問題的分割方法。 [0030] 圖2是說明本發明的第一實施形態的配送計畫的一例圖。 利用圖2來說明有關下車離開型的汽車共享的配送物(車輛)的配送例。在圖2中所謂中心是配送物的配送人員存在，開始配送物的配送之據點(出發據點)。停車場A、停車場B、停車場C是成為配送物的配送源頭或配送去處的據點(配送據點)。配送物的使用者是以預定的預約系統等來進行配送物的預約。使用者是從預約系統輸入配送物的利用台數、利用開始場所(例如停車場B)等的資訊。當使用者所欲從停車場B利用配送物時，若配送物已經存在於停車場B，則利用者可利用該配送物。但，當配送物不存在於停車場B時，配送人員需要從其他的停車場將配送物移動至停車場B。在下車離開型的汽車共享中，使用者是例如若從停車場B往停車場A利用配送物，則將配送物往停車場A就這樣下車離開。於是，可產生多數的使用者利用，例如，配送物遍及於停車場A之類的狀況。中心的配送人員是將遍及的配送物配送至有使用者的需求的停車場A~停車場C。圖2的例子的情況，在停車場A是配送物多餘2台，在停車場B、停車場C是配送物各不足1台。中心的配送人員會將在停車場A多餘的2台的配送物分別各1台配送至停車場B、停車場C，使用者可依照希望利用配送物。圖2是表示符合此條件的配送的實現例。 [0031] 首先，從中心，配送人員2名(k、l)搭乘於1台的配送車1(配送手段)來朝停車場A(1)移動。在停車場A，配送人員k會搭乘於多餘的2台的其中1台的配送物來朝停車場B移動。其他的配送人員l也就這樣搭乘於配送車1來朝停車場B移動(2)。在停車場B，配送人員k會將配送物停止於停車場B，搭乘於配送人員l所駕駛的配送車1。配送人員k、l是從停車場B朝停車場A返回(3)。一旦回到停車場A，則配送人員k是搭乘於多餘的1台的配送物來朝停車場C移動。配送人員l是原封不動搭乘於配送車1來朝停車場C移動(4)。在停車場C，配送人員k會將配送物停止於停車場C，搭乘於配送人員l所駕駛的配送車1。配送人員k、l是從停車場B朝中心返回(5)。若以如此的程序來進行配送，則可符合使用者的要求。在本實施形態中，將如此的狀況的配送物的配送方法定式化為時空網路模型的最小成本流程問題，求取可實行的配送方法之中成本形成最小的方法。 [0032] ＜第1時空網路模型＞ 圖3是說明本發明的第一實施形態的配送計畫的第1時空網路模型的圖。 圖3是將在圖2說明過的配送的實施例予以模型化成時空網路的圖。 圖3的縱軸是表示時間經過，橫軸是表示各據點的場所。圖中，時空上的點是表示各時刻的各據點。圖中，連結2個點的箭號是表示配送物、配送車(配送手段)、配送人員(人)的時空上的移動。各箭號是表示移動源頭據點與移動去處據點，移動所要的時間。實線箭號是表示據點間的移動，二重線箭號是表示停留在同一據點的配送物、配送車、配送人員(移動時間)。附在各箭號來表示的行列的各要素是表示藉由箭號所示的配送來移動的配送物、配送車、配送人員的數量，由上依序表示移動的配送物的數量、移動的配送車的數量、移動的配送人員的人數。例如，實線箭號31的情況，從中心往停車場A，配送物為0台，配送車為1台，配送人員為2名，顯示從時刻t=0到t=1之間移動的情形。二重線箭號32是在停車場A，配送物為2台，配送車為0台，配送人員為0名，顯示從時刻t=0到t=1之間停留的情形。實線箭號33的情況，從停車場A往停車場B，配送物為1台，配送車為1台，配送人員為2名，顯示從時刻t=1到t=2之間移動的情形。二重線箭號34是在停車場A，配送物為1台，配送車為0台，配送人員為0名，顯示從時刻t=1到t=2之間停留的情形。停車場A的配送物從2台變化成1台，是因為配送人員搭乘於2台之中的配送物1台，往停車場B移動所致。有關其他的箭號也同樣。將一個的箭號稱為分支。圖3的分支的集合是對應於在圖2說明的配送計畫者。將配送物配送時，有時在哪個停車場等候配送人員或配達手段等與時間有關的動作會伴隨。在有關配送計畫的既存的數理模型中，大多是以據點為點，以據點間的配送車的移動作為分支來模型化。在本實施形態中，以2次元的時空網路來模型化。藉此，不僅據點間的空間性的移動，還可顯示介入時間的車輛或人的移動。 [0033] ＜根據第1時空網路模型的配送計畫的產生方法＞ 在本實施形態中，求取在圖2、圖3說明那樣的限制時間內符合需要之類的配送計畫配送物中使花在配送的成本最小化的配送計畫。此問題是可作為以下所示的整數計畫問題定式化。 [目的函數] 使花費在配送的配送物、配送手段、配送人員的成本的合計最小化 [限制條件] (1)各據點的流量是符合流量保存規則。 (2)存在於停車場的車輛台數是不超過停車場的停止空間。 (3)在配送手段是出發據點以外一定有配送人員乘坐。 (4)在移動時是在配送物或配送手段一定有配送人員乘坐，移動時的人數是可搭乘於配送物及配送手段的人數的合計以下。 [0034] 為了解開上述的整數計畫問題，配送計畫產生部12是根據初期條件設定部11所受理的初期條件的資訊，產生在圖3所例示般的2次元的時空資訊，產生涉及配送的點之間的分支，以在配送期限內符合各據點的需要數量之方式，產生複數個可將配送物配送至設定有該需要數量的據點之類的分支資訊的集合。然後，配送計畫產生部12從複數的分支資訊的集合之中選擇使成本形成最小的分支資訊的集合。 [0035] 到此為止被提供而來的配送計畫的數理模型或程式是被限於將配送物裝載於卡車或鐵路等的輸送手段來配送的情況。如車輛的配送般，對於包含搭乘輸送的情況，即使是既存技術也有藉由追加限制條件來解開的可能性，但可想像不僅限制條件變複雜，而且手段或路徑的組合數會指數函數性地增大，因此非實用性。若根據本實施形態，則在產生可搭乘輸送的配送物的配送計畫時，藉由作為時空網路模型的最小成本流程問題定式化，可求取可實行的配送方法之中成本最小的配送計畫。 [0036] ＜第2時空網路模型＞ 在第2時空網路模型中，相對於第1時空網路模型，更追加配送據點內的圖表(點資訊及表示配送人員、配送物、配送手段的移動之分支資訊)。藉此，可將配送據點內設為0-1整數計畫問題處理。0-1整數計畫問題是由整數計畫問題來限制變數的幅度，可構成繃緊的緩和問題。藉此，容易放入有效不等式(切割(cut))，謀求計算處理的高速化(計算時間的縮短化)。 [0037] 配送計畫產生部12利用第2時空網路模型來產生配送計畫時，除了在第1時空網路模型的情況說明的時空資訊之外，還針對一個的配送據點，產生：將該配送據點的入口與時刻設成組的點資訊、將該配送據點的出口與時刻設成組的點資訊、對於涉及該配送據點的配送物各1個來將時刻設成組的點資訊。配送計畫產生部12是將涉及入口的點資訊與涉及配送物的點資訊之間的配送人員及配送物的流量的值設定成0或1。配送計畫產生部12是將涉及出口的點資訊與涉及配送物的點資訊之間的配送人員及配送物的流量的值設定成0或1。 [0038] 在此，說明有關配送計畫擬定者朝配送計畫裝置10輸入的參數。在輸入參數是有以下的項目。 亦即，出發據點的集合(Depot)、配送據點的集合(W)、配送期限(dl)、1區間的時間(h)、配送物的種類的集合(P)、配送手段的種類的集合(D)、配送手段的裝載量(cp)、配送物的成本cx(日元/分鐘)、配送手段的成本cy(日元/分鐘)、配送人員的成本cz(日元/分鐘)、移動時間矩陣M(例如，藉由配送手段d之從配送據點w1往w2的移動時間是設為m[d][w1][w2])、各據點的供給數量supply(例如，配送據點w的d的供給數量是設為supply[w、d])、各配送據點的需要數量demand(例如，配送據點w的d的需要數量是設為demand[w、d])。初期條件設定部11是取得該等的參數，作為配送計畫的初期條件設定。 [0039] 作為輸出項目，配送計畫產生部12是將流動於被最適化的配送計畫的時空網路之配送物的流量x((v，s)，(w，t))、配送手段的流量y((v，s)，(w，t))、配送人員的流量z((v，s)，(w，t))輸出至輸出入部13。以((v，s)，(w，t))來表示在時刻s從配送據點v出發，在時刻t到達至配送據點w。在輸出項目是其他有花費在配送的成本等。 [0040] 圖4是說明本發明的第一實施形態的配送計畫的第2時空網路模型的第一圖。 利用圖4來說明第2時空網路。以下，將場所集合設為N，將時間集合設為T，時空網路的圖表G=(V、E)。V是點集合，E是分支集合。點集合V是如以下般定義。 V = {(w，d，p，t)|w∈W，d∈{0}∪P，p∈Swd
，t∈T}在此，Swd
={0，1}(d=0)，Swd
={0，1，...，m}(d≠0)。d=0是表示配送據點的出入口的道路。d=0時，Swd
是取0或1的值，Swd
=0是表示入口，Swd
=1是表示出口。d≠0時，d是表示配送物的種類，Swd
是取0~m的值。m是配送據點w之對於配送物d的供給數量-1或需要數量-1。例如，在配送據點w，配送物a多餘3個時(供給數量=3)，Swd
是取0、1、2的值。例如，在配送據點w，配送物a不足4個時(需要數量=4)，Swd
是取0、1、2、3的值。針對某配送據點w，也將d=0的點稱為道路，將d≠0的點稱為停靠點(port)。停靠點是表示放置1個配送物d的場所。 [0041] 圖4是Depot={0}、W={1,2}、P={a}、T={0,1,2}、S1a
={0}(在配送據點1對於配送物a的供給數量1個)、S2a
={0}(在配送據點2對於配送物a的需要數量1個)的情況的時空網路。其次，利用圖5來說明有關分支集合E。 [0042] 圖5是說明本發明的第一實施形態的配送計畫的第2時空網路模型的第二圖。 以下，將分支集合E設為E=Ex
∪Ey
∪Ez
。Ex
是配送物的分支集合，Ey
是配送手段的分支集合，Ez
是配送人員的分支集合。 配送物的分支集合Ex
是如以下般定義。Ewwx
是表示配送物的據點間的移動之分支的集合、Ewx
是配送物停留在配送據點的道路之分支(等候等)的集合，Ewpx
是表示配送物從道路往停靠點移動之分支的集合，Epwx
是表示配送物從停靠點往道路移動之分支的集合，Epx
是配送物停留在停靠點之分支的集合。 [0043] 配送手段的分支集合Ey
是如以下般定義。Ewwy
是表示配送手段的據點間的移動之分支的集合，Ewy
是配送手段停留在配送據點的道路之分支(等候等)的集合，Ewpy
是表示配送手段從道路往停靠點移動之分支的集合，Epwy
是表示配送手段從停靠點往道路移動之分支的集合。 [0044] 配送人員的分支集合Ez
是如以下般定義。Ewwz
是表示配送人員的據點間的移動之分支的集合，Ewz
是配送人員停留在配送據點的道路之分支(等候等)的集合，Ewpz
是表示配送人員從道路往停靠點移動之分支的集合，Epwz
是表示配送人員從停靠點往道路移動之分支的集合。 [0045] 在圖5顯示以上述定義後的分支集合的表示例。斜方向的實線箭號是表示對應於Ewwx
、Ewwy
、Ewwz
的各集合之分支。縱方向的二重線箭號是表示對應於Ewx
、Ewy
、Ewz
的各集合之分支。橫方向的二點鎖線箭號是表示對應於Ewpx
、Ewpy
、Ewpz
的各集合之分支。斜方向的一點鎖線箭號是表示對應於Epwx
、Epwy
、Epwz
的各集合之分支。縱方向的虛線箭號是表示對應於Epx
的集合之分支。 在第2時空網路模型中，將配送據點內設為0-1整數問題處理，但橫方向的二點鎖線箭號、斜方向的一點鎖線箭號、縱方向的虛線箭號的分支為與此0-1整數問題化關聯，在本實施形態被追加的分支。 [0046] 圖6是說明本發明的第一實施形態的配送計畫的第2時空網路模型的第三圖。 利用圖6來說明有關對於各分支設定的流量向量。如圖6(a)所示般在各分支e是設定有流量向量。 [0047][0048] 此流量向量之中，x[d、e](d∈P，e∈Ex
)是表示配送物的流量，y[d、e](d∈D，e∈Ey
)是表示配送手段的流量，z[e](e∈Ez
)是表示配送人員的流量。舉幾個例子。P={a}，D={車}的情況，流量向量是成為以下般。 [0049][0050] 圖6(b)所示的分支e是表示a為1個，配送車(車)為1台，配送人員(人)為2人的移動。P={a，b}，D={車，摩托車}的情況，流量向量是成為以下般。 [0051][0052] 圖6(c)所示的分支e是表示a為1個，b為1個，配送車(車)為1台，摩托車為1台，配送人員(人)為2人的移動。 [0053] 圖7是說明本發明的第一實施形態的配送計畫的第2時空網路模型的第四圖。 圖7是流量向量為以下般， [0054][0055] 使1個配送物a從配送據點1往配送據點2移動時的時空網路模型。首先，配送人員(人)2人及配送車(車)1台會從Depot出口往配送據點1移動(實線箭號91)。在配送據點1，配送人員1人會朝配送物a的存放處(停靠點0)移動(2點鎖線箭號92)。在配送物a的停靠點0，配送物a會存在至時刻0~1(虛線箭號93)。接著，配送人員1人與配送物a1個會朝配送據點1的出口移動(1點鎖線箭號94)。其次，配送物a1個、配送車1台、配送人員2人會從配送據點1的出口往配送據點2的入口移動(實線箭號95)。其次，配送人員1人與配送物a1個會朝配送據點2的停靠點0移動(2點鎖線箭號96)。接著，配送人員1人會從停靠點0往配送據點2的出口移動(1點鎖線箭號97)。其次，配送人員2人與配送車1台會從配送據點2的出口往Depot入口移動(實線箭號98)。在配送據點2的停靠點0，配送物a會存在至時刻2~3(虛線箭號100)。如圖7所示般，本實施形態是在配送據點1及配送據點2，按每個配送物a1個分配點，且在配送據點的入口、出口各個分配點。為此，配送據點1內及配送據點2內的流量向量的各要素的值是成為0或1。藉此，將配送據點內設為0-1整數計畫問題處理，可使計算時間縮短化。 [0056] 圖8是說明本發明的第一實施形態的配送計畫的第2時空網路模型的第五圖。 圖8是利用第2時空網路模型來表示在圖3所例示的第1時空網路模型者。在圖8中，以更細分化的(1，0，0，0)、(1，a，0，0)、(1，a，1，0)的各列的點集合來表示在圖3中以停車場A的列所示的點集合。有關停車場B、停車場C也同樣。中心、停車場A~停車場C的各者是被賦予對應於出入口的點集合。 [0057] ＜根據第2時空網路模型的配送計畫的產生方法＞ 以上，說明了將配送據點內設為0-1整數計畫問題，作為對於第2時空網路模型的處理的高速化之對策。其次，說明有關配送計畫的產生方法。作為配送手段，針對以配送車及腳踏車配送的情況(亦包含只以配送車配送的情況)、以卡車(裝載配送物)配送的情況的各者進行說明。配送手段的腳踏車是配送人員會乘坐於腳踏車來朝有供給的配送據點移動，在那裡將腳踏車裝載於剩餘的車輛，配送人員駕駛剩餘車輛來朝有需要的配送據點移動這樣的用法。 [0058] 首先，說明有關目的函數。在圖3中，以使成本最小化的情況為例進行說明。在此，使花在配送的移動時間最小化的情況也包含說明。 (以配送車及腳踏車配送的情況) 1.使花在配送的成本最小化的情況的目的函數是對配送物、配送車、配送人員各者的每時間的成本乘上移動時間而加算者。 2.使花在配送的移動時間最小化的情況的目的函數是加算配送物、配送車、配送人員各者的移動時間者。 [0059] (以卡車配送的情況) 1.使花在配送的成本最小化的情況的目的函數是對配送卡車、配送人員各者的每時間的成本乘上移動時間而加算者。 2.使花在配送的移動時間最小化的情況的目的函數是加算配送卡車、配送人員各者的移動時間者。 [0060] 其次，說明有關限制條件。 (以配送車及腳踏車配送的情況與以卡車配送的情況共通) 1.流量保存規則 (1)有關配送物的流量 在配送開始時間點，從存在供給的停靠點出去的流量是成為1。在配送完了時間點，進入存在需要的停靠點的流量會成為1。在其他的點，出去的流量與進入的流量是相等。 [0061] (2)有關配送車的流量 在配送開始時間點，從存在配送車的道路出去的流量是成為配送車的存在台數分。在配送完了時間點，進入存在配送車的道路的流量是成移配送車的存在台數分。在其他的點，出去的流量與進入的流量是相等。 (3)有關配送人員的流量 在配送開始時間點，從存在配送人員的道路出去的流量是成為配送人員的人數分。在配送完了時間點，進入存在配送人員的道路的流量是成為配送人員的人數分。在其他的點，出去的流量與進入的流量是相等。 2.容量限制 配送據點內的配送物、人的移動量為1以下。 [0062] 3.配送據點內的限制1(Ewp
是表示配送物、配送手段、人從道路往停靠點移動之分支的集合) (1)e的去處為d的供給點時 配送物是不會從道路進入停靠點。 (2)e的去處為d的需要點時 配送物與配送人員的流量為取同值(0或1)。 (3)哪個情況皆是腳踏車的台數為配送人員的人數以下。 [0063] 4.配送據點內的限制2(Epw
是表示配送物、配送手段、人從停靠點往道路移動之分支的集合) (1)e的出發點為d的供給點時 配送物與配送人員的流量為取同值(0或1)。 (2)e的出發點為d的需要點時 配送物是不會從停靠點出至道路。 (3)哪個情況皆是腳踏車的台數為配送人員的人數以下。 [0064] 以下，分成以配送車及腳踏車配送的情況及以卡車配送的情況來進行說明。 (以配送車及腳踏車配送的情況) 5.對於從據點往別的據點移動時的分支之限制(Eww
是表示配送物、配送手段、人的據點間的移動之分支的集合) (1)e為腳踏車的分支且車的分支時 有分別在腳踏車及車乘坐配送人員而移動的情況，及將腳踏車裝載於車來移動的情況，限制在車中一定有配送人員乘坐，配送人員的人數為可搭乘人數以下，無法駕駛的腳踏車的台數為可裝載於車的台數以下。 (2)e為腳踏車的分支但不為車的分支時(以腳踏車移動) 腳踏車的台數與人數一致。 [0065] 6.對於在配送據點之中停留的分支之限制 限制在車中一定有人乘坐。 [0066] (以卡車配送的情況) 5.對於從配送據點往別的配送據點移動時的分支之限制 限制在卡車中一定有配送人員乘坐，配送人員的人數為可搭乘人數以下，配送物的總體積為卡車的裝載量以下。 [0067] 6.對於在配送據點之中停留的分支之限制 限制配送物的總體積為卡車的裝載量以下。 [0068] 而且，為了使計算時間縮短化，可適用被稱為切割(cut)的限制式。在整數計畫問題中，將可符合可實行領域的點之不等式稱為有效不等式。整數計畫問題，由於難解，因此首先作為去除整數條件的線形緩和問題處理也多。將追加削減線形緩和問題的解空間之有效不等式的情形稱為切割的追加。藉由切割的追加，線形緩和解接近整數最適解，因此在計算的高速化有強力的效果。藉由切割的追加，可實行領域不被切斷，因此解的最適性被保證。例如，追加規定配送車是1台以上到來之類的切割。於是，可從計算對象排除配送車為3/4台到來非現實性的情況，可提升計算速度。有關切割的具體例是被記載於本案的申請人之日本特願2016-051550的說明書。追加切割的情況，有關在不追加切割的情況下需要數小時以上的整數計畫問題，可以數分鐘解開。 [0069] 配送計畫產生部12是產生以道路(d=0)及停止停靠點(d≠0)來分割配送據點之中的時空網路模型，更利用上述限制條件、被追加的切割來計算分支資訊。若根據本實施形態，則由於可將配送據點內設為0-1整數計畫問題模型化，因此除利用第1時空網路來產生配送計畫時的效果之外，還可取得配送計畫的計算所要的時間可削減的效果。而且，可藉由切割的追加來大幅度縮短計算時間。藉此，例如，比較在各種的初期條件下被產生的配送計畫，可進行更低成本的配送計畫的選擇等，計畫擬定者的便利性會提升。 [0070] 其次，說明有關利用第2時空網路來產生配送計畫的事例。 圖9是表示本發明的第一實施形態的配送問題的一例圖。 圖10是表示對於配送問題來產生的配送計畫的一例圖。 圖9的點c0(depot)是出發據點。其他的點w1~w9是配送據點。例如，在點w7，「d1：2」是表示配送物d1多餘2個的狀態，在點w5，「d2：-1」是表示配送物d2不足1個的狀態。在4點w2，「d1：-1、d2：1」是表示配送物d1不足1個，配送物d2多餘1個的狀態。由圖9所示的配送物d1~d3遍在的狀態，思考配送配送物d1~d3的問題，而使配送物d1~d3在各配送據點不足的狀態能夠消失。以下，分別針對以車及腳踏車配送的情況，以卡車配送的情況，舉配送計畫裝置10所產生的配送計畫的例子。 [0071] 1.以車及腳踏車配送的情況 (1)輸入 ･在depot是配送車1台，腳踏車3台，配送人員3人 ･配送期限：120分鐘 ･配送車的成本：1.5日元/分鐘 ･配送人員的成本：17日元/分鐘 ･d1、d2、d3的成本：1.5日元/分鐘 ･配送車的可搭乘人數：4人、腳踏車的可裝載台數：1台 ･d1的可搭乘人數：1人、腳踏車的可裝載台數：0台 ･d2的可搭乘人數：4人、腳踏車的可裝載台數：1台 ･d3的可搭乘人數：2人、腳踏車的可裝載台數：0台 (2)目的函數 成本最小化 [0072] 若對於此目的函數追加上述的限制條件、切割，而使計算於配送計畫裝置10，則可取得其次般的輸出結果。 (3)輸出 ･以配送車1台、腳踏車1台、配送人員2人，如圖10所示般分成以S1~S7所示的配送路徑及以T1~T9所示的配送路徑的2路徑來配送。 ･配送成本：2768日元 ･必要的時間：60分鐘 如此，可取得在以初期條件設定的配送手段、配送人員、配送期限的條件內使成本最小化的配送計畫。 [0073] 藉由配送計畫裝置10之利用第1時空網路或第2時空網路的配送計畫的產生是在可搭乘輸送的配送物的配送、籌措、售後服務巡迴也可適用。例如，在售後服務中，籌措售後服務所必要的零件等之後須向客戶或巡迴複數的客戶來提供服務。 可將配送物設為使用於售後服務的零件等，將配送主體設為服務人員，將配送手段設為服務人員使用於移動的車輛或零件等的運搬所必要的配送車等，將據點設為客戶或成為售後服務的對象的製品的設置場所，適用上述的數理模型來解開整數計畫問題，服務人員在複數的客戶進行售後服務時，可算出使該巡迴成本或巡迴時間形成最小的巡迴方法(巡迴手段、巡迴路徑)。圖2的情況，例如，只要以停車場A~停車場C(配送據點)作為提供售後服務的客戶，以車輛(配送物)作為零件，以轎車(配送手段)作為服務人員使用於移動的手段、以配送者(配送主體)作為服務人員即可。售後服務的情況，不單只是巡迴客戶，在客戶進行檢查、修理等的作業。只要藉由使用時空網路模型的配送計畫裝置10、10A、10B，便可將服務人員的巡迴行動考慮放進該等作業時間來模型化。 [0074] 上述的例子以外，對於不進行搭乘輸送的配送物的配送也可適用。例如，有被稱為集貨配送(Milk-Run)的集貨方法存在。所謂集貨配送是某製品的製造廠從複數的供應商取得使用於該製品的原材料或零件時，不是使搬入至各供應商，而是製造廠巡迴各供應商來收購原材料等的籌措方法。若藉由集貨配送來進行集貨，則例如藉由以1台的卡車來集貨，與使交納至各供應商的情況作比較，可謀求成本削減、工廠周邊的堵車減輕、環境的負荷減輕。藉由本實施形態的配送計畫裝置10來算出集貨配送的最適的巡迴方法時，例如，若以製品製造廠的訂貨方的工廠等作為有需要汽車共享的配送據點，以原材料或零件的交納業者的工廠等作為車輛多餘的配送據點，則藉由設定與上述「以卡車配送的情況」同樣的目的函數、限制條件等，可算出配送計畫。有關以上述的卡車配送的情況的「配送物的總體積為卡車的裝載量以下」的限制條件是變更成「(原材料的重量×原材料的量+零件的重量×零件數)為卡車的裝載量以下」的限制條件。某供應商的集貨時間範圍被指定時，可藉由在限制條件追加該被指定的集貨時間範圍的資訊來對應。例如，必須使往某供應商的到達時刻形成(從集貨的開始)30分鐘後以後時，只要追加其次般的限制條件，便可產生遵守該供應商的時間範圍限制的集貨計畫。 往某供應商的到達時刻≧30分鐘後 [0075] 到此為止，說明了本實施形態的配送計畫裝置10產生配送計畫的處理。若根據上述的方法，則配送計畫裝置10是可求取配送問題的嚴格的最適解。然而，只靠上述的方法，一旦問題的規模變大(例如，配送據點為20處，配送的車輛為20台)，則會有難以實用性的時間求解的情況。於是，第一區域分割部14是按每個配送區域分割配送據點，將被賦予的配送問題分割成每個分割後的區域的配送問題。其次，說明有關藉由第一區域分割部14的區域分割處理(第一區域分割處理)。 [0076] 圖11是說明本發明的第一實施形態的第一區域分割處理的第一圖。 圖11的左圖是表示在成為配送計畫的對象的地域所存在的配送據點及出發據點的位置。圖中，圓點是表示配送據點(w1~w17)，四角點是表示出發據點(c1~c3)。在配送據點w1~w17是有配送物的需要或剩餘，在出發據點c1~c3是有配送車或配送人員存在。第一區域分割部14是產生使鄰近存在之有需要的配送據點(需要點)與有剩餘的配送據點(供給點)對應的組，更產生集合1個或複數個需要點與供給點的組而分割的一個區域。將藉由第一區域分割部14之區域的產生稱為第一區域分割處理。 [0077] 圖11的右圖是表示第一區域分割部14進行第一區域分割處理的結果。配送據點w1~w5是屬於區域j1，配送據點w6~w9是屬於區域j2，配送據點w10~w17是屬於區域j3。配送計畫產生部12是分別針對分割後的區域j1~區域j3來以上述的方法進行配送計畫的產生。在區域j1~區域j3是分別只含未滿10個的配送據點，為比較的小規模。因此，配送計畫產生部12是可利用上述的時空網路模型來以實用性的時間產生配送計畫。 [0078] 其次，利用圖11來說明第一區域分割處理的概要。首先，第一區域分割部14是將配送據點w1~w17分成需要點、供給點、不為需要點也不為供給點的配送據點。其次，第一區域分割部14是將需要點與供給點附上1對1對應。此時，第一區域分割部14是將需要點與供給點的距離(移動時間)近者彼此間附上對應。例如，在圖11的配送據點w4，配送物d1為多餘1，在配送據點w5，配送物d1為不足1。在配送據點w6，配送物d1為多餘1，在配送據點w7，配送物d1為不足1。在配送據點w8，配送物d2為多餘1，在配送據點w9，配送物d2為不足1。此情況，第一區域分割部14是將需要與供給為一致，距離近的配送據點彼此間附上對應。亦即，第一區域分割部14是將配送據點w4與配送據點w5附上對應(作為組1)，將配送據點w6與配送據點w7附上對應(作為組2)，將配送據點w8與配送據點w9附上對應(作為組3)。有關其他的需要點與供給點也同樣進行附上對應。 [0079] 其次，第一區域分割部14是針對附上對應的需要點與供給點的組，將距離近的組彼此間集合複數來產生1個的區域。此時，若距離近的組無他，則亦可以1個的組作為1個的區域。例如，上面舉的例子時，第一區域分割部14是計算組1的配送據點w4與組2的配送據點w6的距離。然後，若計算後的距離比預定的臨界值小，則將組1及組2分類成同區域，若計算後的距離比預定的臨界值大，則組1及組2是作為別的區域判定。有關組之間的距離的計算，例如，組1與組2的距離的情況，亦可計算配送據點w4與配送據點w7的距離，或分別計算配送據點w4~配送據點w6間、配送據點w4~配送據點w7間、配送據點w5~配送據點w6間、配送據點w5~配送據點w7間的距離，以計算後的距離的平均作為組1與組2的距離。或，亦可以計算後的距離之中最大(或最小)距離作為組1與組2的距離。圖11的例子的情況，第一區域分割部14是例如將組2與組3分類成同區域j2，將組1分類成別的區域j1。第一區域分割部14針對其他的配送據點也進行同樣的處理，按每個區域分類配送據點w1~w17，產生圖11所示的區域j1~區域j3。 [0080] 若藉由第一區域分割處理來產生區域，則配送計畫產生部12是亦可針對區域j1~區域j3的各者，對於以各區域內的各配送據點的配送物的供給數量、需要數量等作為初期條件的配送問題，利用第1時空網路模型或第2時空網路模型來進行配送計畫的產生，但即使不全由初期的狀態來進行利用時空網路模型的計算，亦可以第一區域分割處理時的需要點與供給點的附上對應作為分支資訊利用。亦即，配送計畫產生部12是產生從附上對應的需要點往供給點的分支資訊，藉由使該等接合，可產生符合全部的需要之配送計畫。但，僅連結距離最近的配送據點間的配送路徑，太過於被限定，由移動於複數的據點間的配送計畫全體來看時，有可能遠離被最適化的狀態。於是，第一區域分割部14是進行追加有可能在配送計畫產生部12進行分支資訊的產生時被選擇的配送路徑之處理。然後，配送計畫產生部12是從追加了有可能在連結符合需要及供給的最近的配送據點間的配送路徑所被追加選擇的配送路徑的全部成為候補的配送路徑之中選擇適當的配送路徑來產生分支資訊，產生每個區域的配送計畫。 在圖11中，以需要點、供給點作為配送據點進行說明，但亦可分別為有需要的停靠點、有供給的停靠點。 [0081] 其次，詳細說明有關第一區域分割處理。 圖12是說明本發明的第一實施形態的第一區域分割處理的第二圖。 以下說明本實施形態的第一區域分割處理的處理方式(algorithm)。第一區域分割部14是以以下的程序進行各處理。 1.首先，將各配送據點的各停靠點(第1時空網路模型的情況是配送據點)分成供給點及需要點，產生供給點的集合A、需要點的集合B。設為V1
=A∪B。 2.其次，在V1
追加新的點s、t，設為V=V1
∪{s，t}。 3.從A到B，製作配送路徑e11
~e19
，對於各配送路徑定義根據以該配送路徑所示的路徑來配送時的移動時間之成本函數c(e)。 4.從s到A，從B到t，製作配送路徑，將該等的成本設為0。將從s到A，從B到t的配送路徑的集合設為E2
，將E1
與E2
的和集合設為E。 5.以到此為止的處理來構成2部圖表G=(V，E)。 6.算出G的最小分量最大匹配，亦即成本成為最小之從s往t的路徑。更具體而言，從s往t的路徑之中，從集合A所含的各供給點出來的配送路徑數量為1，進入集合B所含的各需要點的配送路徑的數量為1的條件之下，算出連結s與集合A的各供給點和集合B的各需要點與t之配送路徑的成本的合計成為最小的情況的配送路徑(主問題)。依據4.，從s往集合A的各供給點的成本，從集合B的各需要點往t的成本為0，因此可求取成本成為最小的情況的供給點a1~供給點a3與需要點b1~需要點b3的附上對應方式。在此成本是根據移動時間的函數，因此可求取距離最近的供給點與需要點的1對1匹配。在圖12的例子中，例如連結a1與b1的配送路徑e11
，連結a2與b3的配送路徑e16
，連結a3與b2的配送路徑e18
的組合是成本的合計成為最小的配送路徑，該情況，a1與b1，a2與b3，a3與b2是以1對1附上對應。 7.其次，以被附上對應的2個的停靠點含在同區域的方式，以區域來分割全部的停靠點。此時，若被附上對應的2個的停靠點間的移動時間為預定的值以下，則分類成1個的區域。在圖12的例子中，例如，附上對應的a1與b1、a2與b3、a3與b2之中，以a1與b1作為1個的區域(j4)，以a2與b3、a3與b2作為1個的區域(j5)分類。 [0082] 8.其次，再度作成以1.~5.的程序所作成的2部圖表，產生6.的相對問題(相對問題是可以數學性的操作來產生)。對於分支e=(u，v)，若主問題的最適解X*
(e)為1，則將相對問題的最適解設為Y的情況，成為c(e)-Y*
(u)-Y*
(v)=0(互補性條件：complementary slackness)時，以c^(e)=c(e)-Y*
(u)-Y*
(v)來定義c^(e)，依據0 ≦ c^(e) ≦ ε(ε為預定的定數)來使c^(e)持有寬度。此條件不只是c^(e)=0，c^(e)成為ε以下的情況也意味作為配送路徑的候補處置。亦即，即使在主問題不只是成本形成最小的情況，也可取得持有寬度的解。藉由此解的條件的緩和，在圖12的例子中，例如可重新取得連結a1與b1的配送路徑的候補e19
。 [0083] 以上第一區域分割處理完了。一旦第一區域分割處理完了，則配送計畫產生部12會針對分割後的各區域產生配送計畫。此時，配送計畫產生部12是使用在第一區域分割處理取得的配送路徑的候補來產生符合條件的配送計畫。具體而言，上述的「1.流量保存規則」~「6.對於在配送據點之中停留的分支之限制」之外，還在限制條件加上使用在第一區域分割處理取得的配送路徑的候補(e11
，e16
，e18
，e19
)，計算分支資訊的集合。 [0084] 其次，說明有關本實施形態的配送計畫的產生處理的流程。 圖13是表示本發明的第一實施形態的配送計畫的產生處理的一例的流程圖。 首先，進行配送計畫的配送人員會將配送計畫的初期條件輸入至配送計畫裝置10。輸出入部13是受理該輸入，將受理的資訊輸出至初期條件設定部11。初期條件設定部11是取得配送人員所輸入的初期條件的資訊(步驟S11)。初期條件設定部11是將取得的初期條件的資訊設定於配送計畫的初期條件。所謂初期條件是例如各據點的配送物的供給數量(多餘的數量)、需要數量(不足的數量)、處於出發據點的配送車的台數、配送人員的人數、各據點間的移動時間、配送期限等。 其次，第一區域分割部14是利用在初期條件的資訊所含的各配送據點(或停靠點)的配送物的供給數量、需要數量來進行第一區域分割處理(步驟S12)。有關第一區域分割處理是如利用圖11，圖12來說明般。 [0085] 其次，配送計畫產生部12是按每個在第一區域分割處理取得的區域，以在圖3或圖8所例示的時空網路模型上，符合上述的各限制條件之方式產生分支資訊，產生複數個符合配送期限等的條件之分支資訊的集合(步驟S13)。具體而言，配送計畫產生部12是針對在第一區域分割處理取得的配送路徑的候補，附加配送物的數量、配送車的台數、配送人員的人數的資訊，產生符合各限制條件(「1.流量保存規則」~「6.對於在配送據點之中停留的分支之限制」)的分支資訊。配送計畫產生部12是組合產生的分支資訊，而產生複數個表示至配送期限為止符合各據點的需要台數之類的配送的分支資訊的集合。 [0086] 其次，配送計畫產生部12是按每個區域，按每個產生後的分支資訊的集合來計算總成本(步驟S14)。例如，按每個配送車、配送人員、配送物，每單位時間發生的單位成本會被預先記錄於記憶部15，配送計畫產生部12是對配送車、配送人員、配送物的單位成本乘上各分支所示的時間來計算每個分支的成本(花在配送車、配送人員、配送物的成本的合計)。配送計畫產生部12是計算分支資訊的集合所含的每個分支的成本而將該等合計。合計後的成本為對於1個的分支資訊的集合之成本。配送計畫產生部12是按每個區域，針對全部的分支資訊的集合各者計算成本。 [0087] 其次，配送計畫產生部12是按每個區域，針對計算後的各集合來比較計算後的成本，選擇總成本成為最小的分支資訊的集合(步驟S15)。選擇後的分支資訊的集合是由初期條件所示的出發據點或各配送據點的狀態來表示隨著時間的經過之配送物、配送手段、配送人員的移動(圖3、圖8)。因此，只要根據分支資訊的集合來實行配送，對應於來自使用者的需要之配送便成為可能。亦即，此分支資訊的集合是有關分割後的1個區域之求取的配送計畫。一旦配送計畫產生部12完成產生每個區域的配送計畫，則對於作為初期條件被賦予的全配送據點之配送計畫會被產生。 [0088] 若根據本實施形態，則藉由將在初期條件被賦予的配送據點的全體分割成處於近距離的配送據點的集合(區域)，可縮小配送問題的規模，可削減在各區域的配送計畫的產生處理的計算量。在進行區域分割時算出配送路徑的候補，從該等的候補之中選擇配送路徑來產生分支資訊，因此可更削減計算量。藉此，即使配送物或配送據點的數量多，配送問題形成大規模的情況，也可以實用性的時間(例如10分鐘)產生配送計畫。 配送計畫產生部12是亦可在步驟S13中不使用配送路徑的候補來產生分支資訊，產生配送計畫。 亦可加進區域的配送計畫的成本來再構成區域。例如，有對於配送成本的差顯著偏離的區域的一對，藉由擴大儘快配送完了的區域，縮小很晚配送完了的區域，可降低全體的成本的情況。 [0089] ＜第二實施形態＞ 其次，參照圖14~圖18來說明針對大規模的配送問題以實用性的時間產生配送計畫的其他的方法(第二實施形態)的配送計畫系統。在第一實施形態是著眼於比較近距離存在的需要點及供給點，集合一對的需要點及供給點來產生的小規模的區域。在第二實施形態是以出發據點為中心進行區域分割處理，在分割後的各區域內產生配送計畫。將第二實施形態的區域分割處理稱為第二區域分割處理。 [0090] 圖14是表示本發明的第二實施形態的配送計畫系統的一例的機能方塊圖。 本發明的第二實施形態的構成之中，與本發明的第一實施形態之構成配送計畫裝置10的機能部相同者是附上同樣的符號，省略各個的說明。第二實施形態的配送計畫裝置10A是取代第一實施形態的構成的第一區域分割部14，而具備第二區域分割部16。 第二區域分割部16是取得複數個表示預先被設定的配送路徑及以該配送路徑配達時的成本之單位路徑資訊，該預先被設定的配送路徑是從出發據點出發來進行一部分的配送據點間的配送而回到原來的出發據點。第二區域分割部16是從複數的單位路徑資訊的組合之中，選擇該組合所含之配送據點的數量成為預定的數量以內，且成本成為最小的組合。此單位路徑資訊的組合所含的出發據點及配送據點的集合為分割後的區域。 第二區域分割部16是藉由配送計畫裝置10A所具備的CPU從記憶部15讀出程式實行而實現。 [0091] 圖15是說明本發明的第二實施形態的配送問題的第二區域分割處理的第一圖。 在圖15中，圓點是表示配送據點w1~配送據點w17，四角點是表示出發據點c1~出發據點c3。在本實施形態中，首先，列舉可思考的所有的需要･供給的模式(pattern)之後，按每個出發據點作成多數個符合各模式的需要･供給之單純的配送路徑。在此，在出發據點c1~出發據點c3是存在充分量的配送車(卡車)或配送人員，配送據點w1，w3，w6，w8，w10為供給點，配送據點w2，w5，w7，w9，w11為需要點。涉及配送的配送物的種類是設為相同。此情況，所謂符合每個出發據點的需要･供給之單純的配送路徑是例如從出發據點c1出發到配送據點w1(R1)，在配送據點w1取配送物，朝配送據點w2運該配送物(R2)，一旦配達終了，則回到出發據點c1(R3)的配送路徑(作為配送路徑1)。同樣，從出發據點c1經由配送據點w3、配送據點w5來回到出發據點c1的路徑(作為配送路徑2)，從出發據點c1經由配送據點w6、配送據點w7來回到出發據點c1的配送路徑(作為配送路徑3)，從出發據點c2經由配送據點w8、配送據點w9來回到出發據點c2的配送路徑(作為配送路徑4)，從出發據點c3經由配送據點w10、配送據點w11來回到出發據點c3的配送路徑(作為配送路徑5)等。該等的單純的路徑是預先被作成多數個，被記錄於記憶部15。雖未圖示，但例如以下般的路徑也被作成預先被記錄於記憶部15。從出發據點c1經由配送據點w1及配送據點w5來回到出發據點c1的配送路徑，從出發據點c1經由配送據點w1及配送據點w7來回到出發據點c1的配送路徑，從出發據點c1經由配送據點w3及配送據點w2來回到出發據點c1的配送路徑，從出發據點c1經由配送據點w10及配送據點w11來回到出發據點c1的配送路徑，從出發據點c2經由配送據點w1及配送據點w2來回到出發據點c2的配送路徑，從出發據點c3經由配送據點w1及配送據點w2來回到出發據點c3的配送路徑等。 [0092] 所謂單純的配送路徑，不是單指經由1個的供給點、1個的需要點來回到原來的出發據點的配送路徑(巡迴2個的配送據點而回到原來的出發據點的路徑)，亦可為從出發據點c1出發，在配送據點w1取2個配送物，將配送物各一個送達配送據點w2及配送據點w4，回到出發據點c1的配送路徑(配送據點為3個的例子)。或，亦可為從出發據點c1出發，在配送據點w1取1個配送物，送達配送據點w2，接著在配送據點w3取1個配送物，送達配送據點w4，回到出發據點c1的配送路徑(配送據點為4個的例子)。 [0093] 圖16是說明本發明的第二實施形態的配送問題的第二區域分割處理的第二圖。 如圖16所圖示般，在記憶部15是記錄有複數個預先被準備的單純的配送路徑。對單純的配送路徑附上對應，記錄以該配送路徑配送時的成本。成本是例如作為移動時間的函數被賦予。預先被記錄的單純的配送路徑與成本的資訊是越多越可產生更精度高(接近嚴格的最適解)的配送計畫。將包含圖16所例示的單純的配送路徑及對應於該配送路徑的成本之資訊稱為單位路徑資訊。 [0094] 圖17是說明本發明的第二實施形態的配送問題的第二區域分割處理的第三圖。 其次，利用圖15~圖17來說明有關第二區域分割部16之第二區域分割處理。 前提是在記憶部15記錄有在圖16所例示的單位路徑資訊。有關配送的要求事項(每個配送據點的需要數量、供給數量、出發據點的配送車的台數、配送人員的人數等)的初期條件會被賦予。 1.首先，第二區域分割部16是從記憶部15讀出在初期條件被賦予的配送要求(需要)之中，包含符合一部分的配送路徑之單位路徑資訊。 2.其次，第二區域分割部16是組合讀出的單位路徑資訊所含的配送路徑，作成符合全部的配送要求的路徑，作為暫定解。在此，作為組合預先被記錄的單純的配送路徑，求取符合全部的要求且成本成為最便宜的組合的方法，雖可使用被稱為集合分割途徑的數學的手法，氮此手法的情況，一旦單純的配送路徑的數量形成膨大，則該組合的最適化問題是有無法以實用性的時間解開的情況。為此，第二區域分割部16是記憶部15所記憶的多數的單純的配送路徑之中，使用少數的配送路徑來產生初期解，利用被稱為列產生法的數學的手法來一邊追加單純的配送路徑，一邊求取最適解。用以產生初期解的少數的單純的配送路徑是可用任意的方法來選擇。在初期解的產生是可使用一般被提供的解算器(solver)。此時，在1個的出發據點所擔負的配送據點的數量設置限制(例如，包含需要點與供給點，6據點以內等)。在此，之所以在配送據點的數量設置限制(上限)，是為了削減計算量，使處理高速化。有關限制的配送據點的數量是例如亦可實際進行計算，將可在實用性的時間內求解的情況的配送據點的數量定為上限。初期解，例如可取得圖15所示的解(配送路徑1~配送路徑5)。 [0095] 3.其次，第二區域分割部16是藉由列產生法(周知的數學的手法)來從未選擇的單純的配送路徑之中選擇1個或複數的配送路徑，計算更換已被選擇的某單純的配送路徑與選擇後的單純的配送路徑時的減少成本。所謂減少成本是從對應於藉由更換而作為重新解被選擇的單純的配送路徑來記錄的成本(圖16的成本欄)的合計減去對應於更換前被選擇的原本的單純的配送路徑來記錄的成本的合計之值。 [0096] 4.第二區域分割部16是若在3.計算後的減少成本為負的值(更換後較便宜)，則以在3.選擇的單純的配送路徑來更新既存的全配送路徑的一部分。例如，有關朝配送據點w7的配送，在圖15所示的初期解，雖配送路徑3會被選擇，但第二區域分割部16是取代此配送路徑3，計算更換成從出發據點c2經由配送據點w6、配送據點w7來返回至出發據點c2的路徑(設為配送路徑6)時的減少成本。依據圖16，配送路徑3的成本是2500，配送路徑6的成本是1500，所以減少成本是-1000。因此，第二區域分割部16是以配送路徑6來更新配送路徑3。將更新後的單純的配送路徑的組合顯示於圖17。計算減少成本的情況也第二區域分割部16是以1個的出發據點所擔負的據點的數量成為限制內的方式選擇配送路徑。 [0097] 此例是舉將單純的配送路徑從出發據點c1的擔負範圍移往出發據點c3的擔負範圍的例子，但配送路徑的更新並非限於此例。例如，亦可為在圖15的狀態，將出發據點c1所擔負的路徑內的配送路徑1與配送路徑2的組合更新成從出發據點c1經由配送據點w3、配送據點w2來朝出發據點c1返回的路徑與從出發據點c1經由配送據點w1、配送據點w5來朝出發據點c1返回的路徑的組合之方法。或，亦可為在配送據點w1有2個作為對象的配送物，將配送路徑1與配送路徑2的組合更新成從出發據點c1經由配送據點w1、配送據點w2、配送據點w3、配送據點w5來朝出發據點c1返回的配送路徑之方法。 [0098] 5.第二區域分割部16是重複3.~4.的處理，在可減少成本的路徑變無的時間點終了第二區域分割處理。一旦第二區域分割處理終了，則按每個出發據點，限制條件(例如6據點)以內的配送據點會被附上關聯。此被附上關聯的複數的配送據點與出發據點的集合(例如區域j6~區域j8的各者)為藉由第二區域分割處理所取得的區域。 [0099] 其次，說明有關本實施形態的配送計畫的產生處理的流程。 圖18是表示本發明的第二實施形態的配送計畫的產生處理的一例的流程圖。有關與圖13同樣的處理是簡單地進行說明。 首先，初期條件設定部11是取得配送人員等所輸入的初期條件的資訊(步驟S11)。 其次，第二區域分割部16是利用初期條件的資訊所含之在各據點的配送物的供給數量、需要數量、在出發據點的配送人員數、配送車的種類及數量來進行第二區域分割處理(步驟S121)。有關第二區域分割處理是如利用圖15~圖17來說明般。 [0100] 其次，配送計畫產生部12是按每個在第二區域分割處理所取得的區域(在第二區域分割處理所取得的出發據點及被該出發據點附上關聯的複數的配送據點的集合)，在圖3或圖8所例示的時空網路模型上，以符合上述的各限制條件之方式產生分支資訊，產生複數個符合配送期限等的條件之分支資訊的集合(步驟S131)。在圖17的例子中，配送計畫產生部12是針對區域j6~區域j8的各者來產生複數個分支資訊的集合。配送計畫產生部12是亦可無關在第二區域分割處理使用的單純的配送路徑及其組合，進行分支資訊的產生，或亦可利用單純的配送路徑及其組合來進行分支資訊的產生。 [0101] 其次，配送計畫產生部12是按每個產生後的分支資訊的集合來計算總成本(步驟S14)。在圖17的例子中，配送計畫產生部12是針對區域j6~區域j8的各者來計算每個產生後的分支資訊的集合的總成本。 其次，配送計畫產生部12是比較計算後的總成本，選擇總成本成為最小的分支資訊的集合(步驟S15)。在圖17的例子中，配送計畫產生部12是針對區域j6~區域j8的各者來選擇總成本成為最小的分支資訊的集合。 選擇後的分支資訊的集合是表示各區域的配送計畫。亦即，一旦配送計畫產生部12完成產生每個區域的配送計畫，則有關分割前的全配送據點的配送計畫會被產生。 [0102] 若根據本實施形態，則可藉由將在初期條件被賦予的有需要或供給的全部的配送據點一個一個對出發據點附上關聯，可分割成每個出發據點的配送區域。藉由分割成小規模的配送區域，按每個該等的配送區域進行配送計畫的產生處理，相較於以在初期條件被賦予的全部的有需要的配送據點作為對象來產生配送計畫的情況，可削減配送計畫的產生所必要的計算量。藉此，即使配送物或配送據點的數量多，配送問題形成大規模時，也可以實用性的時間(例如10分鐘)擬定配送計畫。 [0103] ＜第三實施形態＞ 更參照圖19~圖22來說明有關針對大規模的配送問題以實用性的時間產生配送計畫的其他的方法(第三實施形態)的配送計畫系統。第一實施形態、第二實施形態是根據初期條件所含的每個配送據點的需要及供給的資訊來空間性地分割(區域分割)配送問題，藉此分割成小規模的配送問題，藉由解開小規模的配送問題來實現處理的高速化之方法。此第三實施形態是根據初期條件所含的配送物的配送限制時間的資訊來藉由時間性的分割使配送問題小問題化，按分割後的各時間產生配送計畫。 [0104] 圖19是表示本發明的第三實施形態的配送計畫系統的一例的機能方塊圖。 本發明的第三實施形態的構成之中，與本發明的第一實施形態之構成配送計畫裝置10B的機能部相同者是附上同樣的符號，省略各個的說明。第三實施形態的配送計畫裝置10B是取代第一實施形態的構成的第一區域分割部14，而具備時間分割部17。配送計畫裝置10B是取代配送計畫產生部12，而具備配送計畫產生部12a。 時間分割部17是將在初期條件被賦予的配送限制時間分割成複數的時間，藉此分割成每個分割配送問題的時間之規模小的配送問題。將時間分割部17所分割而形成的各時間稱為區間。 配送計畫產生部12a是按照時間分割部17所分割而形成的區間，將設定了不同的目的函數或限制的配送問題解開而產生配送計畫。例如，配送計畫產生部12a是產生從分割後的各區間的最初的時刻的配送物的配送狀況(前1個的區間終了時的配送的進展狀況)開始配送，在該區間內配送物盡可能更多被配送至有需要的配送據點之類的配送計畫。時間分割部17有關分割而形成的區間之中最後的區間是產生在其最後的區間內配送物會被配送至在初期條件所示之有需要的配送據點之中成為未配送的配送據點的全部之配送計畫。 時間分割部17、配送計畫產生部12a是藉由配送計畫裝置10B所具備的CPU從記憶部15讀出程式實行而實現。 [0105] 圖20是說明本發明的第三實施形態的配送問題的時間分割處理的第一圖。 圖20的左圖是表示某配送問題的初期狀態。在配送據點w1是配送物d1多餘2個，在配送據點w2是配送物d1多餘3個。在配送據點w3是配送物d1不足4個，在配送據點w4是配送物d1不足1個。可思考由此初期狀態將配送物d1配送成為在120分鐘以內符合配送據點w3、配送據點w4的需要之問題。在第一實施形態、第二實施形態中，按每個區域分割配送據點w1~配送據點w4。在本實施形態中，分割作為初期條件被賦予的配送限制時間120分鐘。 具體而言，時間分割部17是將配送限制時間120分鐘分割成複數的時間(區間)。例如，時間分割部17是將配送限制時間120分鐘分割成前半60分鐘、後半60分鐘的2個的區間。時間分割部17所分割而形成的1區間的時間的長度或區間的數量可為任意。例如，時間分割部17是亦可將120分鐘分割成90分鐘及30分鐘，或分類成每40分鐘的3個的區間。時間分割部17是亦可以比在初期條件被賦予的配送限制時間更只短預定的時間的時間作為分割後的第1個的區間，以剩下的時間作為最後的區間。 [0106] 圖20的右圖是表示時間分割部17將配送限制時間120分鐘分割成各60分鐘的2個區間。一旦時間分割部17進行配送限制時間的分割，則配送計畫產生部12a會針對分割而形成的第1個的區間進行配送計畫的產生。此時，配送計畫產生部12a不是以符合全部的要求為目標，而是產生在1個目的區間的終了時間點盡可能使符合多的要求之類的配送計畫。所謂以符合全部的要求為目標，此例的情況，是以60分鐘配送配送物d1，使能滿足配送據點w3及配送據點w4的需要數量。所謂盡可能使符合多的要求之類的配送計畫，例如在開始配送之後60分鐘後，在配送據點w3的配送物d1的不足為3個，在配送據點w4的配送物d1的不足為0個的配送計畫，及在配送據點w3的配送物d1的不足為2個，在配送據點w4的配送物d1的不足為0個的配送計畫產生時，意指後者的配送計畫。 [0107] 其次，說明有關針對第1個的區間設定的目的函數、限制。其次說明的目的函數等較一般是時間分割部17會將配送限制時間分割成N個的情況，除了最後的區間(第N個的區間)，針對全部的區間(沿著時間的流程，從第1個的區間到N-第1個的區間)使用的目的函數等。 [0108] (目的函數) 使以下的(1)~(7)的和最小化。 (1)據點間的移動成本 可抑制據點間的移動成本。 (2)符合需要供給的時刻 在促使盡可能加快符合需要供給的時刻下，解的選擇項會減少(儘管是即使在後的時刻也可同樣的配送的情況，藉由只解最先的配送，可減少選擇項，連帶計算量的削減)。 (3)在符合需要的情況進行供給的停靠點號碼 在促使由小的號碼的停靠點進行供給下，解的選擇項會減少(儘管是即使配送至大的號碼的停靠點也無問題的情況，藉由選定成最小的號碼的停靠點，可減少解的選擇項，削減計算量)。 (4)存在於配送據點的人、配送手段之中，在之後的區間不使用者的數量(g[w]) 促使在計算中的區間的最後的時刻，使在其次的區間未被利用於配送物的配送之多餘的人･配送手段移動至出發據點。例如，在第一區間的最後的時刻，可搭載腳踏車而移動的配送手段變無時，在第二區間，由於不將腳踏車使用於配送，因此使移動至出發據點。 (5)符合需要點的需要、供給點的供給的情況，存在於該等的配送據點的人、配送手段的數量 促使在符合需要供給的配送據點，人或配送手段不會殘留。在本實施形態中，有關進行時間分割的各個的區間(雖繼承前一個的區間的配送狀況)，但獨立算出配送計畫，因此有可能在之後的區間不使用的多餘的配送人員或配送手段殘留於配送據點。於是，追加(4)、(5)來使多餘的配送人員等不會殘留於配送據點。 (6)在供給點剩下的配送物的數量、在需要點不足的配送物的數量 (7)配送物不足的需要點的數量 藉由(6)、(7)，可促使在該區間的最後的時刻，盡可能儘量地多符合需要供給的條件。 目的函數是作為使對應於各項目(1)~(7)的各個的內容之函數的和最小化的情形表示，但亦可對目的函數的各項(各項目的函數)乘算任意的係數來進行加權。 [0109] 其次，說明有關限制。 1.流量保存規則 (I)分割後的各區間之中最後的區間以外的區間的情況 (1)有關配送物的流量 在區間的開始時間點，從存在供給的停靠點出去的流量是成為1，從不存在供給的停靠點出去的流量是成為0。在其他的點，不是區間的終了時間點的情況，出去的流量與進入的流量是相等。 (2)有關配送車的流量 在區間的開始時間點，從道路出去的流量是成為存在於該據點的配送車的數量。在其他的點，不是區間的終了時間點的情況，出去的流量與進入的流量是相等。 (3)有關配送人員的流量 在區間的開始時間點，從道路出去的流量是成移存在於該據點的配送人員的數量。在其他的點，不是區間的終了時間點的情況，出去的流量與進入的流量是相等。 [0110] (Ⅱ)分割後的各區間之中最後的區間的情況 (1)有關配送物的流量 在區間的開始時間點，從存在供給的停靠點出去的流量是成為1，從不存在供給的停靠點出去的流量是成為0。在區間的終了時間點，進入至存在需要的停靠點的流量是成為1，進入至不存在需要的停靠點的流量是成為0。在其他的點，出去的流量與進入的流量是相等。 (2)有關配送車的流量 在區間的配送開始時間點，從道路出去的流量是成為存在於該據點的配送車的數量。在區間的終了時間點，從道路進入的流量是成為在該據點有需要的配送車的數量。在其他的點，出去的流量與進入的流量是相等。 (3)有關配送人員的流量 在區間的配送開始時間點，從道路出去的流量是成為存在於該據點的配送人員的人數。在區間的終了時間點，從道路進入的流量是成為在該據點有需要的配送人員的人數。在其他的點，出去的流量與進入的流量是相等。 [0111] 2.容量限制 配送據點內的配送物、人的移動量為1以下。 [0112] 3.配送據點內的限制1(Ewp
是表示配送物、配送手段、人從道路往停靠點移動之分支的集合) (1)e的去處為d的供給點時 配送物是不從道路進入停靠點。 (2)e的去處為d的需要點時 配送物與配送人員的流量為取同值(0或1)。 (3)哪個情況皆是腳踏車的台數為配送人員的人數以下。 [0113] 4.配送據點內的限制2(Epw
是表示配送物、配送手段、人從停靠點往道路移動之分支的集合) (1)e的出發點為d的供給點時 配送物與配送人員的流量為取同值(0或1)。 (2)e的出發點為d的需要點時 配送物是不從停靠點出去至道路。 (3)哪個情況皆是腳踏車的台數為配送人員的人數以下。 [0114] 5.對於有關區間的最後的時刻的罰則(penalty)的變數之限制 對於任意的道路w，將在其次的區間不使用的人或配送手段的數量設為g[w]時，放入g[w] ≦ f[w]的限制。 f[w]是表示任意的數。藉由調整f[w]的值，可變更能容許殘存於區間的最後的情形的人、配送手段的數量。 [0115] 6.對於從道路往道路的分支之限制 (1)從道路往道路的分支為車的分支時(載腳踏車的可能性有) 限制在配送車中一定有配送人員乘坐，配送人員的人數為可搭乘人數以下，無法駕駛的腳踏車的台數為可裝載於車的台數以下。 (2)從道路往道路的分支為腳踏車的分支且為徒步的分支，不為車的分支時(以腳踏車或徒步移動) 腳踏車台數與以徒步移動的人數的和是相等於移動該分支的配送人員的人數。 (3)從道路往道路的分支為腳踏車的分支，但不為車的分支，也不為徒步的分支時(以腳踏車移動) 腳踏車台數與移動該分支的配送人員的人數相等。 (4)從道路往道路的分支為徒步的分支，但不為車的分支，也不為腳踏車的分支時(以徒步移動) 以徒步移動的人數與移動該分支的配送人員的人數相等。 [0116] 7.對於在配送據點之中停留的分支之限制 限制在車中一定有人乘坐，腳踏車無法路上停車，配送物的總體積為卡車的裝載量以下。 [0117] 配送計畫產生部12a是利用第1或第2時空網路來解開在該等的目的函數或限制條件下被定式化的配送問題(整數計畫問題)。在實際的計算中，亦可加上切割來謀求更高速化。在圖20的右圖表示配送計畫產生部12a針對第1個的區間來產生配送計畫的結果的一例。若根據此計畫，則開始配送之後60分鐘後，在配送據點w1及配送據點w2，配送物d1多餘1個，在配送據點w3，配送物d1不足2個，在配送據點w4，可知配送物d1的不足解消。 [0118] 如此分割時間來計算的方法的優點，可舉其次般的點。首先，藉由將配送限制時間從120分鐘形成60分鐘，可減少計算所必要的參數的數量，削減計算量。在圖20中，基於說明的方便起見，舉配送據點、配送物的數量皆少的例子，但在計算在120分鐘以內配送20個以上的配送物至20處的配送據點時，例如，可知變數為41000，甚至限制為61000。相對於此，以前半的60分鐘來劃分時間，產生上述盡可能符合多的要求之類的配送計畫時，例如，可將變數的數量壓到14000，將限制的數量壓到18000程度。藉由如此減少參數的數量，可削減計算量，使處理高速化。其次，說明有關後半的60分鐘的配送計畫的產生處理。 [0119] 圖21是說明本發明的第三實施形態的配送問題的時間分割處理的第二圖。 在圖21顯示使用分割後的後半60分鐘來進行的配送計畫的產生處理。整理有關後半60分鐘的開始時間點的各配送據點的需要數量與供給數量。在後半的開始時間點的配送據點w1及配送據點w2，配送物d1多餘1個，在配送據點w3，配送物d1不足2個。配送計畫產生部12a必須使用後半60分鐘來產生的配送計畫是將配送據點w1及配送據點w2的配送物d1配送成為在剩下的60分鐘以內符合配送據點w3的需要之問題。有關配送據點w4，因為配送據點w4的不足是在前半60分鐘解消，所以不含在後半60分鐘的配送問題中。 [0120] 其次，說明有關針對以將配送限制時間分割成N個時的最後的區間作為對象的配送問題設定的目的函數、限制。此情況的目的函數或限制等是與利用第1時空網路模型(圖3)、第2時空網路模型(圖8)來說明者同樣。亦即，配送計畫產生部12a是產生以至最後的區間的最後的時刻為止符合全部的需要之方式完成配送配送物的配送計畫之中，使成本或配送所花的移動時間最小化的配送計畫。 [0121] 若舉有關後半的60分鐘(最後的區間)的處理的優點，則首先與前半60分鐘同樣，可舉藉由將配送限制時間形成60分鐘，可減少參數的數量，削減計算量，藉此可謀求處理的高速化的點。進一步，可舉藉由在至最後的區間為止的各區間所被算出盡可能多符合要求之類的配送計畫，剩下的需要會減少，因此必須解開的配送問題的規模會變小。例如，在圖20、圖21所示的例子中，不單只是各需要點的需要數量減少，符合在配送據點w4的需要，因此成就了減少配送據點的數量。藉此更可縮小配送問題的規模。將在120分鐘以內配送20個以上的配送物至20處的配送據點的問題分割成每個60分鐘時，在後半60分鐘的區間，例如，可使變數的數量減少至8000，使限制的數量減少至12000程度。如此，若根據本實施形態，則藉由時間的分割之參數的減少，及時間上先的區間中產生的配送計畫，由於可使在之後的區間的配送問題的規模縮小化，因此可使配送計畫的產生處理高速化。在第三實施形態中，由於不進行區域分割，因此與第一實施形態、第二實施形態作比較，可仍舊確保配送據點分布的大域性來產生配送計畫，可產生更被最適化的配送計畫。 [0122] 其次，說明有關本實施形態的配送計畫的產生處理的流程。 圖22是表示本發明的第三實施形態的配送計畫的產生處理的一例的流程圖。有關與圖13、圖18同樣的處理是簡單地進行說明。 首先，進行配送計畫的配送人員會將配送計畫的初期條件輸入至配送計畫裝置10B。輸出入部13是受理該輸入，將受理的資訊輸出至初期條件設定部11。初期條件設定部11是取得配送人員所輸入的初期條件的資訊(步驟S11)。 其次，時間分割部17是利用初期條件的資訊所含的配送限制時間來進行時間分割處理(步驟S122)。有關時間分割處理是如利用圖20來說明般。例如，時間分割部17是將初期條件的配送限制時間分割成2~3個。或，以經驗法則等對於有需要的配送據點之配送完了為可以一定的比例估計的時間(例如75分鐘)得知時，藉由使用者的設定，時間分割部17是亦可以該時間單位(例如75分鐘)來分割原來的配送限制時間，以剩下的時間作為最後的區間。或，時間分割部17是亦可以比在初期條件被賦予的配送限制時間更只短預定的時間的時間作為分割後的第1個的區間，以剩下的時間作為最後的區間。其次，配送計畫產生部12a是對於被分割的各區間依時間順序產生配送計畫。首先，配送計畫產生部12a是對計數器(counter)變數n設定1(步驟S123)。其次，配送計畫產生部12a是設定第n個的區間的配送問題(步驟S124)。例如，若為第1個的區間，則配送計畫產生部12a是將初期條件的資訊所含的各配送據點的需要數量、供給數量、存在於出發據點的配送手段的數量、配送人員的人數、第1個的區間的時間利用於配送限制時間，盡可能符合多的要求，使根據成本最便宜的配送計畫的產生用的目的函數、限制條件之整數計畫問題定式化。 [0123] 其次，配送計畫產生部12a是對於定式化的整數計畫問題，利用在圖3或圖8所例示的第1或第2時空網路模型，產生複數個符合各限制，且盡可能滿足各配送據點的需要之類的分支資訊的集合(步驟S132)。其次，配送計畫產生部12a是按每個產生後的分支資訊的集合來計算總成本(步驟S14)。其次，配送計畫產生部12a是比較計算後的總成本，選擇總成本成為最小的分支資訊的集合(步驟S15)。藉此，對於第n個(此次是第1個)的區間之配送計畫會被產生。 [0124] 其次，配送計畫產生部12a是判定n是否與N相等(步驟S16)。n與N相等時(步驟S16；Yes)，相對於分割後的全部的區間之配送計畫會被產生，因此終了配送計畫的產生處理。n不與N相等時(步驟S16；No)，配送計畫產生部12a是對n加算1(步驟S17)，重複來自步驟S124的處理。 [0125] 具體而言，例如，加算1後的n的值為2時，配送計畫產生部12a是在步驟S124中，將實行針對第1個的區間產生的配送計畫後的結果的各配送據點的需要數量、供給數量、存在於出發據點的配送手段的數量、配送人員的人數、第2個的區間的時間予以利用在配送限制時間，使整數計畫問題定式化。此時，配送計畫產生部12a是若n與N相等，則使符合全部的要求的配送計畫的產生用的目的函數、限制之整數計畫問題定式化(與第一實施形態、第二實施形態同樣的目的函數等)。當n不與N相等時，配送計畫產生部12a是使盡可能符合多的要求的配送計畫的產生用的目的函數、限制之整數計畫問題定式化。配送計畫產生部12a是解開整數計畫問題來產生第n個的區間的配送計畫。配送計畫產生部12a是針對全部的區間，至產生配送計畫為止(n與N相等為止)重複步驟S124~步驟S17的處理。針對第1個的區間~第N個(最後)的區間的各者來依序產生的配送計畫的全體為符合在初期條件被賦予的要求的配送計畫。 [0126] 在圖22的流程圖中，將最初在初期條件被賦予的配送限制時間分割成幾個的區間，例如，亦可其次般構成配送計畫處理。 1.首先，設定比在初期條件被賦予的配送限制時間短的區間(第一配送限制時間)，產生在該區間內盡可能多且可便宜配送的配送計畫。 2.其次，針對在該第一配送限制時間內無法完成配送的配送物，更設定接續於第一配送限制時間的預定長度的區間(第二配送限制時間)，產生在設定的區間內盡可能配送多的配送物之配送計畫。 3.離配送開始的區間的長度的合計不會超過配送限制時間的期間，重複1.~2.的步驟。亦即，一邊延長時間，一邊找到解(時間展開法)。最初的區間或延長的區間的長度是可按照配送狀況來任意設定。 4.離配送開始的時間的合計超過配送限制時間時，以其次的區間作為最後，產生在最後的區間內(以之前剛產生的配送計畫的完了時間點作為開始時刻，以第一配送限制時間的開始時刻為基準，以初期條件所含的配送限制時間經過的時刻作為終了時刻的時間內)符合全部的需要的成本成為最小之類的配送計畫。 [0127] 實際進行本實施形態的配送計畫產生處理時，對於配送據點為20處，配送物為20個程度的配送問題，可以實用性的時間(10分鐘以內)算出準最適解(與嚴格的最適解作比較，目的函數的值的差為10%以下)。 [0128] 本實施形態的配送計畫產生方法是亦可利用在以下那樣的場面。例如，必須在120分鐘將全部的車輛配送至有需要的停車場時，首先針對前半60分鐘產生配送計畫。一旦完成產生配送計畫，則實際根據該配送計畫，開始車輛的配送。進行車輛的配送的期間，產生剩下的60分鐘的配送計畫。如此，藉由並行配送計畫的產生處理及配送，例如在進行汽車共享的服務的現場，可有效地使用時間。 [0129] 上述的配送計畫裝置10、10A、10B的各處理的過程是以程式的形式來記憶於電腦可讀取的記錄媒體，藉由配送計畫系統的電腦讀出此程式實行來進行上述處理。在此所謂電腦可讀取的記錄媒體是意指磁蝶、光磁碟、CD-ROM、DVD-ROM、半導體記憶體等。亦可藉由通訊線路來將此電腦程式配訊至電腦，接受此配訊後的電腦實行該程式。 [0130] 上述程式是亦可為用以實現前述的機能的一部分者。更亦可以將前述的機能與已經被記錄於電腦系統的程式的組合來實現者，所謂的差分檔案(差分程式)。配送計畫裝置10、10A、10B是亦可以1台的電腦來構成，或以可通訊連接的複數的電腦所構成。 [0131] 其他，在不脫離本發明的主旨的範圍，可適當將上述實施形態的構成要素置換成周知的構成要素。此發明的技術範圍並非限於上述的實施形態，可在不脫離本發明的主旨範圍內施加各種的變更。 [0132] 配送人員是配送主體的一例，配送車、卡車、腳踏車是配送手段的一例，汽車共享的共用的車輛是配送物的一例。第一區域分割部14、第二區域分割部16、時間分割部17分別為分割部的一例。 [產業上的利用可能性] [0133] 若根據上述的配送計畫系統、配送計畫方法及程式，則可擬定一種以實用性的時間使相對於大規模的配送問題的成本或移動時間最小化之配送計畫。[0026] <First Embodiment> Hereinafter, a delivery planning system according to an embodiment of the present invention will be described with reference to FIGS. 1 to 13. Fig. 1 is a functional block diagram showing an example of the delivery planning system of the first embodiment of the present invention. In this embodiment, the delivery planning system is constituted by, for example, a computer device such as a PC or a server device. The computer device is composed of arithmetic unit such as CPU (Central Processing Unit), memory unit such as ROM (Read Only Memory), RAM (Random Access Memory), HDD (Hard Disk Drive), and other hardware such as network interface. . [0027] The delivery planning device 10 of FIG. 1 is an example of a delivery planning system. The delivery planning device 10 is a device that calculates a delivery means, a delivery route, etc. that minimize the cost for a delivery plan including a ride-on transportation. In the present embodiment, a scene in which a vehicle shared by users is delivered to a place where the user starts to use it in a drop-off-and-away car sharing is taken as an example to draw up an optimal delivery plan for the delivery. In the case of delivering a delivery, for example, it is required to select a delivery method and a delivery route that minimize the cost. For the formulation of the distribution plan to minimize the cost formation, various methods have been provided so far. However, the delivery of car-sharing vehicles, for example, is different from the delivery of goods such as express delivery. That is the point where people can board the vehicle when delivering the vehicle. For example, in each of the bases of the base A, the base B, the base C, the base D, and the base E, there is a state of surplus or insufficient vehicles. In such a situation, in order to move a vehicle from a base where the vehicle is surplus to a base where the vehicle is insufficient, for example, the following methods can be considered in response to the needs of the user. (1) A one-person delivery person patrols the bases with trucks that can load vehicles, loads the excess vehicles on the trucks, and delivers them to the deficient bases. (2) A plurality of delivery personnel board the delivery car and move to each base. Part of the delivery personnel in the delivery vehicle transfer to the remaining vehicle once they arrive at the base where the vehicle is redundant, and drive and deliver the vehicle to the base where the vehicle is insufficient (board transportation). In such a situation, it is not easy to know which delivery method is best for delivery, and what route is used for delivery, so as to minimize the cost. The delivery planning device 10 of the present embodiment introduces mathematical models or restrictions based on mathematical insights into delivery plans when transportation is possible, thereby providing a high-speed and efficient delivery plan that minimizes costs. method. The delivery planning device 10 of the present embodiment generates a delivery plan for delivering a delivery to a delivery site, based on the required number of deliveries (insufficient number) for each delivery site that is given as the initial condition of the delivery. And the supply quantity (excess quantity) to divide the total distribution problem indicated by the initial conditions into partial small-scale distribution problems. In this way, even if the number of delivery vehicles or bases is large, it is possible to draw up a delivery plan in a practical time (for example, 10 minutes). [0028] As shown in FIG. 1, the delivery planning device 10 includes an initial condition setting unit 11, a delivery plan generation unit 12, an input/output unit 13, a first area division unit 14, and a storage unit 15. The initial condition setting unit 11 is a delivery vehicle (e.g., a vehicle used by a user), a delivery subject, a movable delivery, or a delivery means of the delivery subject (e.g., a delivery vehicle such as a truck) that can be boarded and moved by the delivery subject (e.g., delivery personnel) ) The required quantity and supply quantity of deliveries at each delivery location at any one of the places where one stays, and one or more departure locations that indicate the initial positions of the delivery entity and the delivery means (such as those of companies that provide car sharing services) Use the information of the base), the available delivery means and the information of the delivery subject at the departure base, and the information of the delivery deadline, as the initial condition setting in generating the delivery plan. The related parameters will be explained in detail later. The delivery plan generation unit 12 is to calculate point information and branch information, and generate at least one set of branch information that will deliver the required number of deliveries within the delivery period to the delivery base set with the required number. This point information It is to set the delivery base and the departure base and the time based on the start of the delivery as a set. The branch information is between the two point information involved in the delivery in the point information, which indicates the delivery involved in the delivery, the delivery subject and the delivery Means of flow. The departure base and delivery base are collectively called bases. The input/output unit 13 accepts user input operations. The input/output unit 13 outputs information of the delivery plan based on the set of branch information generated by the delivery plan generation unit 12 to a display or the like. The first area dividing unit 14 is to make a demand point included in a group of distribution sites where there is a required delivery, and one supply point included in a group of delivery sites that is the supply source of the delivery, from the one demand point The movement time to the one supply point will be minimized, and the correspondence will be attached, resulting in a set of combinations of the corresponding demand points and the supply points. That is, the first area dividing unit 14 divides the total distribution problems of all the distribution site groups that require delivery based on the initial conditions into the distribution problems of the distribution area units shown in the generated collection (from the distribution in the collection). Including demand point group to supply point group distribution problem). Then, the delivery plan generation unit 12 generates a set of branch information for each of the divided small-scale delivery problems. The storage unit 15 stores information necessary for the generation of the distribution plan. The initial condition setting unit 11, the delivery plan generation unit 12, and the first area division unit 14 are implemented by, for example, a CPU (Central Processing Unit) included in the delivery planning device 10 reading programs from the memory unit 15 achieve. [0029] The delivery planning device 10 uses a spatio-temporal network model to generate a delivery plan for a given delivery problem. First, explain about the spatio-temporal network model and the generation method of the distribution plan using the spatio-temporal network model, and then explain the division method of the distribution problem when the relevant problem becomes large-scale. [0030] FIG. 2 is a diagram illustrating an example of a delivery plan according to the first embodiment of the present invention. An example of delivery (vehicle) for a drop-off type car sharing delivery item (vehicle) will be explained using FIG. 2. The so-called center in FIG. 2 is a base (departure base) where the delivery personnel of the delivered items exist and the delivery of the delivered items is started. The parking lot A, the parking lot B, and the parking lot C are bases (delivery bases) that become the source of delivery or destination of the delivery. The user of the delivery makes a reservation for the delivery using a predetermined reservation system or the like. The user inputs information such as the number of used delivery units and the use start location (for example, parking lot B) from the reservation system. When the user wants to use the delivery from the parking lot B, if the delivery already exists in the parking lot B, the user can use the delivery. However, when the delivery does not exist in the parking lot B, the delivery person needs to move the delivery to the parking lot B from another parking lot. In a drop-off and leave-type car sharing, for example, if the user uses a delivery item from the parking lot B to the parking lot A, the delivery item will be dropped off and left in the parking lot A. Therefore, it can be used by a large number of users, for example, a situation where the delivery is all over the parking lot A. The delivery personnel in the center deliver all the deliveries to the parking lot A to the parking lot C where the users demand. In the case of the example in FIG. 2, there are two more deliveries in the parking lot A, and there are less than one delivery each in the parking lot B and the parking lot C. The delivery staff of the center will deliver the two excess deliveries in parking lot A to parking lot B and parking lot C, respectively, and users can use the delivery as desired. Figure 2 shows an implementation example of delivery that meets this condition. [0031] First, from the center, two delivery personnel (k, 1) board a delivery vehicle 1 (delivery means) to move toward the parking lot A(1). In the parking lot A, the delivery person k will board the delivery item of one of the two excess vehicles and move to the parking lot B. The other delivery personnel 1 just get on the delivery vehicle 1 and move toward the parking lot B (2). In parking lot B, delivery person k will stop the delivery at parking lot B and board the delivery vehicle 1 driven by delivery person l. The delivery personnel k and l return from the parking lot B to the parking lot A (3). Once back to the parking lot A, the delivery person k rides on the excess delivery and moves to the parking lot C. The delivery person 1 rides in the delivery vehicle 1 as it is to move toward the parking lot C (4). In the parking lot C, the delivery person k will stop the delivery at the parking lot C and board the delivery vehicle 1 driven by the delivery person l. The delivery personnel k and l return from the parking lot B to the center (5). If such a procedure is used for distribution, it can meet the requirements of users. In this embodiment, the delivery method of the delivery in such a situation is formulated as a minimum cost process problem of the spatiotemporal network model, and the method with the smallest cost formation among the feasible delivery methods is determined. [0032] <The first spatio-temporal network model> FIG. 3 is a diagram illustrating the first spatio-temporal network model of the delivery plan of the first embodiment of the present invention. Fig. 3 is a diagram showing the example of the delivery described in Fig. 2 modeled into a spatiotemporal network. The vertical axis of FIG. 3 represents the passage of time, and the horizontal axis represents the location of each base. In the figure, the points in time and space represent the bases at each time. In the figure, the arrow connecting two points represents the movement in time and space of a delivery item, a delivery vehicle (delivery means), and a delivery person (person). Each arrow indicates the movement source base and the movement destination base, and the time required to move. The solid arrow indicates the movement between bases, and the double-line arrow indicates the delivery items, delivery vehicles, and delivery personnel staying at the same base (movement time). The elements attached to the rows indicated by the arrows indicate the number of deliveries, delivery vehicles, and delivery personnel that are moved by the delivery indicated by the arrows, and the number of deliveries moved and moved from the top The number of delivery vehicles, the number of moving delivery personnel. For example, in the case of the solid arrow 31, from the center to the parking lot A, there are 0 delivery items, 1 delivery vehicle, and 2 delivery personnel, showing the situation of moving from time t=0 to t=1. The double line arrow 32 is in the parking lot A, with 2 delivery items, 0 delivery vehicles, and 0 delivery personnel, showing the staying situation from time t=0 to t=1. In the case of a solid arrow 33, from parking lot A to parking lot B, there is one delivery item, one delivery vehicle, and two delivery personnel, showing the movement from time t=1 to t=2. The double line arrow 34 is in the parking lot A, with 1 delivery item, 0 delivery vehicles, and 0 delivery personnel, showing the staying situation from time t=1 to t=2. The change of the delivery items in the parking lot A from two to one is because the delivery personnel boarded one of the two delivery items and moved to the parking lot B. The same goes for other arrows. Call an arrow a branch. The set of branches in FIG. 3 corresponds to the delivery planner described in FIG. 2. When delivering deliveries, there may be time-related actions such as waiting for delivery personnel or delivery methods in which parking lot. Most of the existing mathematical models related to delivery plans are based on bases and the movement of delivery vehicles between bases as branches. In this embodiment, a two-dimensional space-time network is modeled. In this way, not only the spatial movement between the bases, but also the movement of vehicles or people involved in the time can be displayed. [0033] <Generation method of delivery plan based on the first spatio-temporal network model> In this embodiment, the delivery plan that meets the needs within the limited time described in FIGS. 2 and 3 is obtained. A distribution plan that minimizes the cost of distribution. This problem can be formulated as the integer planning problem shown below. [Purpose function] Minimize the total cost of delivery items, delivery methods, and delivery personnel spent on delivery [Conditions] (1) The flow of each site is in compliance with the flow preservation rules. (2) The number of vehicles in the parking lot does not exceed the parking space of the parking lot. (3) If the delivery means is outside the departure base, there must be a delivery person on board. (4) When moving, there must be a delivery person on the delivery or delivery means, and the number of people when moving is less than the total number of people who can board the delivery and delivery means. [0034] In order to solve the above integer plan problem, the delivery plan generation unit 12 generates the two-dimensional space-time information as illustrated in FIG. 3 based on the initial condition information received by the initial condition setting unit 11, and generates related The branch between the delivery points generates a collection of multiple branch information such as those that can deliver the delivery to the locations where the required number is set in a manner that meets the required quantity of each base within the delivery period. Then, the delivery plan generation unit 12 selects the set of branch information that minimizes the cost from among the sets of plural branch information. [0035] The mathematical model or program of the delivery plan provided so far is limited to the case where the delivery is loaded on a truck or railroad transportation means for delivery. Like the delivery of vehicles, even the existing technology may be solved by adding restriction conditions for situations involving transportation. However, it is conceivable that not only the restriction conditions will become complicated, but the number of combinations of means or routes will be exponentially functional. Land increases, so it is not practical. According to this embodiment, when generating a distribution plan for a delivery that can be transported by a ride, by formulating the minimum cost process problem as a spatio-temporal network model, it is possible to obtain the least costly distribution among the feasible distribution methods plan. [0036] <The second spatio-temporal network model> In the second spatio-temporal network model, compared to the first spatio-temporal network model, a graph in the distribution site (point information and information indicating the delivery personnel, delivery items, and delivery means Mobile branch information). In this way, the distribution site can be set as an integer of 0-1 to plan the problem processing. The 0-1 integer plan problem is the integer plan problem to limit the range of variables, which can constitute a tight alleviation problem. Thereby, it is easy to put in the effective inequality (cut), and speed up the calculation process (shorten the calculation time). [0037] When the delivery plan generation unit 12 uses the second spatio-temporal network model to generate the delivery plan, in addition to the spatio-temporal information described in the first spatio-temporal network model, it also generates: The point information where the entry and time of the delivery site are set as a group, the point information where the exit and the time of the delivery site are set as a group, and the point information that sets the time as a group for each one of the items related to the delivery site. The delivery plan generation unit 12 sets the value of the flow of delivery personnel and the delivery between the point information related to the entrance and the point information related to the delivery to 0 or 1. The delivery plan generating unit 12 sets the value of the flow rate of delivery personnel and the delivery between the point information related to the exit and the point information related to the delivery to 0 or 1. [0038] Here, the parameters input to the delivery planning device 10 by the delivery plan planner will be described. There are the following items in the input parameter. That is, the collection of departure locations (Depot), the collection of delivery locations (W), the delivery date (dl), the time of one section (h), the collection of types of delivery (P), and the collection of types of delivery means ( D), the loading capacity of the delivery means (cp), the cost of the delivery cx (yen/minute), the cost of the delivery means cy (yen/minute), the cost of the delivery staff cz (yen/minute), the movement time Matrix M (for example, the travel time from the delivery site w1 to w2 by the delivery means d is set to m[d][w1][w2]), the supply quantity of each site supply (for example, the delivery site w's d The supply quantity is set to supply[w, d]), and the required quantity demand of each delivery site (for example, the required quantity of d in the delivery site w is set to demand[w, d]). The initial condition setting unit 11 acquires these parameters and sets them as initial conditions of the delivery plan. [0039] As an output item, the delivery plan generation unit 12 is the flow x ((v, s), (w, t)) and delivery means of the delivery items flowing in the space-time network of the optimized delivery plan The flow rate y ((v, s), (w, t)) of, and the flow rate z ((v, s), (w, t)) of the delivery personnel are output to the input/output unit 13. Let ((v, s), (w, t)) denote that it departs from the delivery site v at time s and arrives at the delivery site w at time t. The output items are other costs spent on delivery. [0040] FIG. 4 is a first diagram illustrating a second spatiotemporal network model of the delivery plan according to the first embodiment of the present invention. Use Figure 4 to illustrate the second space-time network. In the following, set the place set to N, the time set to T, and the graph of the spatiotemporal network G=(V, E). V is the point set, and E is the branch set. The point set V is defined as follows. V = {(w,d,p,t)|w∈W,d∈{0}∪P,p∈S wd ,t∈T} Here, S wd ={0,1}(d=0) , S wd = {0, 1,..., m} (d≠0). d=0 indicates the road of the entrance and exit of the delivery base. When d=0, S wd is a value of 0 or 1, S wd =0 means entrance, and S wd =1 means exit. When d≠0, d represents the type of delivery, and S wd takes a value from 0 to m. m is the supply quantity -1 or the required quantity -1 for the delivery d at the delivery site w. For example, when there are more than three deliveries a at the delivery site w (supply quantity=3), S wd takes values of 0, 1, and 2. For example, when there are less than 4 deliveries a at the delivery site w (required quantity=4), S wd takes values of 0, 1, 2, and 3. For a certain delivery location w, the point where d=0 is also called a road, and the point where d≠0 is called a port. A stop is a place where one delivery d is placed. [0041] FIG. 4 is Depot={0}, W={1,2}, P={a}, T={0,1,2}, S 1a ={0} (for delivery items at delivery site 1 The supply quantity of a is 1), S 2a = {0} (the required quantity of delivery a at the distribution site 2 is 1). Next, use FIG. 5 to explain the branch set E. [0042] FIG. 5 is a second diagram illustrating the second spatiotemporal network model of the delivery plan of the first embodiment of the present invention. Below, set the branch set E as E=E x ∪E y ∪E z . E x is the branch set of delivery items, E y is the branch set of delivery means, and E z is the branch set of delivery personnel. The branch set E x of the distribution is defined as follows. E wwx is a collection of branches that represent the movement between the delivery points, E wx is a collection of branches (waiting, etc.) where the delivery stays on the road of the delivery location, and E wpx is the branch that represents the movement of the delivery from the road to the stop collection, E pwx is a distribution branch were to move the collection from the road stops, E px is the distribution was stuck in a branch stops the collection. [0043] The branch set E y of the delivery means is defined as follows. E wwy is a collection of branches that represent the movement of the delivery means between the bases, E wy is a collection of branches (waiting, etc.) on the road where the delivery means stays at the delivery location, and E wpy is the branch that means the delivery means moves from the road to the stop Epwy is a collection of branches that represent the movement of the delivery means from the stop to the road. [0044] The branch set E z of delivery personnel is defined as follows. E wwz is a collection of branches that represent the movement of the delivery personnel between locations, E wz is a collection of branches (waiting, etc.) on the road where the delivery personnel stay at the delivery location, and E wpz is the branch that represents the delivery personnel moving from the road to the stop The set of Epwz is the set of branches that the delivery personnel move from the stop to the road. [0045] FIG. 5 shows an example of the branch set defined above. The solid arrow in the oblique direction represents the branch corresponding to each set of E wwx , E wwy , and E wwz . The double-line arrow in the vertical direction represents the branch corresponding to each set of E wx , E wy , and E wz . The horizontal two-point lock line arrow indicates the branch corresponding to each set of E wpx , E wpy , and E wpz . The one-point lock arrow in the oblique direction represents the branch corresponding to each set of Epwx , Epwy , and Epwz . The dashed arrow in the vertical direction represents the branch of the set corresponding to Epx . In the second spatio-temporal network model, the distribution point is set to 0-1 integer problem, but the two-point lock line arrow in the horizontal direction, the one-point lock line arrow in the oblique direction, and the dashed arrow in the vertical direction branch into and This 0-1 integer problem is related to the branch added in this embodiment. [0046] FIG. 6 is a third diagram illustrating the second spatiotemporal network model of the delivery plan of the first embodiment of the present invention. The flow vector set for each branch will be explained using FIG. 6. As shown in Fig. 6(a), a flow vector is set in each branch e. [0047] [0048] In this flow vector, x[d, e] (d ∈ P, e ∈ E x ) is the flow of delivery, and y [d, e] (d ∈ D, e ∈ E y ) is The flow of delivery means, z[e](e∈E z ) is the flow of delivery personnel. Give a few examples. In the case of P={a} and D={car}, the flow vector becomes as follows. [0049] [0050] The branch e shown in FIG. 6(b) represents a movement in which a is one, a delivery vehicle (car) is one, and a delivery person (person) is two persons. In the case of P={a, b} and D={car, motorcycle}, the flow vector is as follows. [0051] [0052] The branch e shown in FIG. 6(c) represents a movement in which a is one, b is one, a delivery vehicle (vehicle) is one, a motorcycle is one, and a delivery person (person) is two persons. . [0053] FIG. 7 is a fourth diagram illustrating the second spatiotemporal network model of the delivery plan of the first embodiment of the present invention. Figure 7 is the flow vector as follows, [0054] [0055] A spatiotemporal network model when one delivery a is moved from the delivery site 1 to the delivery site 2. First, 2 delivery personnel (persons) and 1 delivery vehicle (car) will move from the Depot exit to the delivery site 1 (solid arrow 91). At delivery site 1, one delivery person will move toward the storage place (stop 0) of delivery a (two-point lock line arrow 92). At stop 0 of delivery a, delivery a will exist until time 0~1 (dashed arrow 93). Next, a delivery person and a delivery item a1 will move toward the exit of the delivery site 1 (1-dot chain arrow 94). Next, 1 delivery item a, 1 delivery vehicle, and 2 delivery personnel will move from the exit of delivery site 1 to the entrance of delivery site 2 (solid arrow 95). Secondly, 1 delivery person and delivery a1 will move to stop point 0 of delivery base 2 (two-point lock line arrow 96). Then, one delivery person will move from stop 0 to the exit of delivery base 2 (1 point lock line arrow 97). Secondly, 2 delivery personnel and 1 delivery vehicle will move from the exit of delivery site 2 to the entrance of Depot (solid arrow 98). At stop 0 of delivery location 2, delivery a will exist until time 2~3 (dashed arrow 100). As shown in FIG. 7, in this embodiment, there are 1 distribution points for each delivery a at the distribution site 1 and the distribution site 2, and the distribution points are at the entrance and exit of the distribution site. Therefore, the value of each element of the flow vector in the delivery site 1 and the delivery site 2 becomes 0 or 1. In this way, it is possible to shorten the calculation time by setting the distribution site as an integer of 0-1 to plan the problem processing. [0056] FIG. 8 is a fifth diagram illustrating the second spatiotemporal network model of the delivery plan according to the first embodiment of the present invention. FIG. 8 shows the use of the second spatiotemporal network model to represent the first spatiotemporal network model illustrated in FIG. 3. In Fig. 8, the points in each column of (1, 0, 0, 0), (1, a, 0, 0), and (1, a, 1, 0) are more subdivided as shown in Fig. 3 The points shown in the column of parking lot A are gathered in. The same applies to parking lot B and parking lot C. Each of the center and parking lot A to parking lot C is a set of points assigned to the entrance and exit. [0057] <Method of generating a distribution plan based on the second spatio-temporal network model> Above, the problem of setting the distribution site as an integer of 0-1 was explained as the speeding up of processing for the second spatio-temporal network model The countermeasures. Next, explain how to generate the distribution plan. As a delivery method, the following describes the case of delivery by delivery vehicle and bicycle (including delivery only by delivery vehicle), and the case of delivery by truck (loading delivery). The bicycle of the delivery method is a usage in which the delivery person rides on the bicycle to move to the delivery base where there is supply, where the bicycle is loaded on the remaining vehicles, and the delivery person drives the remaining vehicles to move to the delivery base in need. [0058] First, the purpose function is explained. In FIG. 3, a case where the cost is minimized is taken as an example for description. Here, the case of minimizing the movement time spent on delivery is also included. (In the case of delivery by delivery vehicles and bicycles) 1. The objective function for minimizing the cost of delivery is the cost per time of the delivery, delivery vehicle, and delivery personnel multiplied by the travel time. 2. The objective function in the case of minimizing the travel time spent on delivery is to add the travel time of the delivery items, delivery vehicles, and delivery personnel. [0059] (In the case of delivery by truck) 1. The objective function for minimizing the cost of delivery is the addition of the cost per time of the delivery truck and the delivery personnel multiplied by the travel time. 2. The objective function for minimizing the travel time spent on delivery is to add the travel time of the delivery truck and delivery personnel. [0060] Next, the relevant restrictions are explained. (The case of delivery by delivery vehicles and bicycles is the same as the case of delivery by trucks) 1. Flow preservation rules (1) The flow of a delivery item is 1 at the time of the start of delivery. At the point in time when the delivery is completed, the traffic entering the order where there is a need will become 1. At other points, the outgoing flow is equal to the incoming flow. [0061] (2) Regarding the flow of delivery vehicles, at the time of the start of the delivery, the flow out of the road where the delivery vehicles exist is divided into the number of existing delivery vehicles. At the point in time when the delivery is completed, the traffic entering the road where the delivery vehicle exists is the number of minutes of the existing delivery vehicle. At other points, the outgoing flow is equal to the incoming flow. (3) The flow of related delivery personnel At the point of the delivery start time, the flow out of the road where the delivery personnel exist is divided by the number of delivery personnel. At the point in time when the delivery is completed, the traffic entering the road where the delivery personnel exist is divided by the number of delivery personnel. At other points, the outgoing flow is equal to the incoming flow. 2. The capacity limit is 1 or less for delivery items and people moving within the delivery site. [0062] 3. Restrictions within the distribution site 1 (E wp is the collection of the branch that represents the delivery, the delivery means, and the person moving from the road to the stop) (1) When the destination of e is the supply point of d, the delivery will not enter the stop from the road. (2) When the destination of e is the required point of d, the flow rate of the delivery and the delivery personnel is the same value (0 or 1). (3) In all cases, the number of bicycles is less than the number of delivery personnel. [0063] 4. Restrictions within the distribution site 2 (E pw is the collection of delivery items, delivery means, and branches that move people from the stop to the road) (1) When the starting point of e is the supply point of d, the flow of the delivery and the delivery personnel is the same value (0 or 1 ). (2) When the starting point of e is the required point of d, the delivery will not leave the stop point to the road. (3) In all cases, the number of bicycles is less than the number of delivery personnel. [0064] The following description is divided into the case of delivery by a delivery vehicle and bicycle and the case of delivery by a truck. (In the case of delivery by delivery vehicle and bicycle) 5. Restrictions on branches when moving from a base to another base (E ww is a collection of branches that represent the movement of delivery items, delivery means, and people's bases) (1) e is the branch of bicycles, and the branch of the car may be moved by the delivery personnel on the bicycle and the car respectively, and When a bicycle is mounted on a car to move, it is restricted to a delivery person in the car, the number of delivery personnel is less than the number of passengers, and the number of bicycles that cannot be driven is less than the number that can be loaded on the car. (2) When e is a bicycle branch but not a bicycle branch (moving by bicycle) The number of bicycles matches the number of people. [0065] 6. Restrictions on the branches staying in the delivery site restrict people in the vehicle. [0066] (In the case of delivery by truck) 5. The restriction on the branch when moving from the delivery site to another delivery site is that there must be delivery personnel in the truck, and the number of delivery personnel is less than the number of passengers. The total volume is less than the load capacity of the truck. [0067] 6. Restrictions on the branches staying in the delivery site restrict the total volume of the delivery to less than the load of the truck. [0068] Furthermore, in order to shorten the calculation time, a restriction expression called cut may be applied. In the integer project problem, the inequality of the point that can meet the feasible field is called the effective inequality. Since the integer planning problem is difficult to solve, it is first dealt with as a linear relaxation problem that removes integer conditions. The addition of the effective inequality that reduces the solution space of the linear relaxation problem is called the addition of cutting. With the addition of cutting, the linear relaxation solution is close to the integer optimal solution, so it has a powerful effect in speeding up the calculation. With the addition of cutting, the practicable area is not cut, so the optimal solution is guaranteed. For example, it is added that the delivery vehicle is a cut of more than one arrival. Therefore, the unrealistic situation of 3/4 delivery vehicles arriving can be excluded from the calculation object, and the calculation speed can be increased. A specific example of cutting is described in the specification of Japanese Patent Application 2016-051550 of the applicant of this case. In the case of additional cutting, the integer plan that takes more than a few hours without additional cutting can be solved in a few minutes. [0069] The delivery plan generation unit 12 generates a spatio-temporal network model that divides the distribution bases by roads (d=0) and stops (d≠0), and uses the above restrictions and additional cuts. Calculate branch information. According to this embodiment, it is possible to model the 0-1 integer plan problem in the distribution site, so in addition to the effect of using the first spatiotemporal network to produce the distribution plan, the distribution plan can also be obtained The effect of reducing the time required for the calculation. Moreover, the calculation time can be greatly reduced by adding cutting. By this, for example, compared with the distribution plans generated under various initial conditions, it is possible to select a lower-cost distribution plan, etc., and the convenience of plan planners will be improved. [0070] Next, an example of using the second spatiotemporal network to generate a distribution plan is described. Fig. 9 is a diagram showing an example of a delivery problem in the first embodiment of the present invention. Fig. 10 is a diagram showing an example of a delivery plan for delivery problems. The point c0 (depot) in Fig. 9 is the starting point. The other points w1~w9 are distribution bases. For example, at point w7, "d1: 2" means that there are more than two deliveries d1, and at point w5, "d2: -1" means that there are less than one delivery d2. At 4 o'clock w2, "d1: -1, d2: 1" means that the delivery item d1 is less than one, and the delivery item d2 is more than one. From the state in which the delivery items d1 to d3 are present in FIG. 9, the problem of the delivery items d1 to d3 is considered, so that the state of the delivery items d1 to d3 being insufficient at each delivery site can disappear. Hereinafter, examples of the delivery plan generated by the delivery planning device 10 are given for the case of delivery by car and bicycle, and the case of delivery by truck. [0071] 1. Delivery by car and bicycle (1) Input･In the depot, there are 1 delivery car, 3 bicycles, and 3 delivery personnel. Delivery deadline: 120 minutes. Delivery car cost: 1.5 yen/minute ･Cost of delivery personnel: 17 yen/minute ･Cost of d1, d2, d3: 1.5 yen/minute ･The number of people that can be carried on the delivery vehicle: 4 people, the number of bicycles that can be loaded: 1 ･Available for d1 Number of people: 1 person, the number of bicycles that can be loaded: 0 units ･The number of people that can be loaded on d2: 4 people, the number of bicycles that can be loaded: 1 ･The number of people that can be used for d3: 2 people, the number of bicycles that can be loaded: 0 unit (2) objective function cost minimization [0072] If the above-mentioned restriction conditions and cuts are added to this objective function and calculated in the delivery planning device 10, the next general output result can be obtained. (3) Output ･With 1 delivery vehicle, 1 bicycle, and 2 delivery personnel, as shown in Figure 10, it is divided into the delivery route shown by S1~S7 and the delivery route shown by T1~T9. Distribution. ･Delivery cost: 2,768 yen ･Required time: 60 minutes. In this way, you can obtain a delivery plan that minimizes costs within the conditions of delivery means, delivery personnel, and delivery deadlines set in the initial conditions. [0073] The generation of a distribution plan using the first spatiotemporal network or the second spatiotemporal network by the distribution planning device 10 is also applicable to the distribution, financing, and after-sales service rounds of deliveries that can be transported by boarding. For example, in after-sales service, after procuring necessary parts for after-sales service, etc., it is necessary to provide services to customers or customers who have patrolled the numbers. The delivery can be set as parts for after-sales service, the delivery body can be set as the service personnel, and the delivery means can be set as the delivery vehicles necessary for the transportation of vehicles or parts, etc. used by the service personnel. For customers or products that are the subject of after-sales service, apply the above mathematical model to solve the problem of integer planning. Service personnel can calculate the tour cost or tour time when performing after-sales services for multiple customers. The smallest tour method (tour means, tour path). In the case of Figure 2, for example, as long as parking lot A ~ parking lot C (delivery base) are used as customers to provide after-sales service, vehicles (delivery items) are used as parts, and cars (delivery means) are used as the means for service personnel to move. The delivery person (delivery subject) can be used as the service staff. In the case of after-sales service, it is not only to visit customers, but to perform inspections and repairs at customers. As long as the distribution planning devices 10, 10A, and 10B using the spatio-temporal network model are used, the touring operations of the service personnel can be modeled by considering the operation time. [0074] In addition to the above examples, it is also applicable to the delivery of deliveries that are not transported on board. For example, there is a collection method called Milk-Run. The so-called collection and delivery refers to a method of raising raw materials or parts used in the product from multiple suppliers. Instead of moving in to each supplier, the manufacturer patrols each supplier to purchase raw materials. If the collection is carried out by consolidating and delivery, for example, by consolidating the goods by one truck, compared with the case of delivering to each supplier, it is possible to reduce costs, reduce traffic jams around the factory, and load the environment. Ease. When using the delivery planning device 10 of the present embodiment to calculate the optimal round trip method for collection and delivery, for example, if the ordering factory of a product manufacturer is used as a delivery base where car sharing is required, the delivery of raw materials or parts The manufacturer’s factory, etc., is used as a delivery base for redundant vehicles, and the delivery plan can be calculated by setting the same purpose function and restriction conditions as the above-mentioned "delivery by truck". Regarding the above-mentioned truck delivery, the restriction condition of "the total volume of the delivery is less than the load of the truck" is changed to "(the weight of raw materials × the amount of raw materials + the weight of parts × the number of parts) is the load of the truck Below" restrictions. When the collection time range of a certain supplier is specified, it can be matched by adding the information of the specified collection time range to the restriction conditions. For example, when the arrival time to a certain supplier must be 30 minutes later (from the start of the collection), only the next few restrictions can be added to produce a collection plan that complies with the time limit of the supplier. The arrival time to a certain supplier ≧30 minutes later [0075] So far, the process of generating a delivery plan by the delivery planning device 10 of this embodiment has been described. According to the above-mentioned method, the delivery planning device 10 can obtain a strict optimal solution to the delivery problem. However, with the above-mentioned method alone, once the scale of the problem becomes larger (for example, there are 20 delivery locations and 20 delivery vehicles), it may be difficult to solve practically in time. Therefore, the first area dividing unit 14 divides the distribution site for each distribution area, and divides the assigned distribution problem into the distribution problem for each divided area. Next, a description will be given of the area division processing (first area division processing) by the first area division unit 14. [0076] FIG. 11 is a first diagram illustrating the first region division processing in the first embodiment of the present invention. The left diagram of FIG. 11 shows the locations of delivery bases and departure bases existing in the area targeted by the delivery plan. In the figure, the dots represent the delivery locations (w1~w17), and the four corners represent the departure locations (c1~c3). At the delivery locations w1~w17, there is a need or surplus of delivery items, and there are delivery vehicles or delivery personnel at the departure locations c1~c3. The first area dividing unit 14 generates a group corresponding to the necessary delivery sites (needs) existing nearby and the remaining delivery sites (supply points), and also generates a group of one or more demand points and supply points. And a divided area. The generation of regions by the first region dividing unit 14 is referred to as a first region dividing process. [0077] The right diagram of FIG. 11 shows the result of the first area dividing process performed by the first area dividing unit 14. Delivery locations w1~w5 belong to area j1, delivery locations w6~w9 belong to area j2, and delivery locations w10~w17 belong to area j3. The delivery plan generation unit 12 generates delivery plans in the above-mentioned method for the divided areas j1 to j3, respectively. In area j1 to area j3, there are only less than 10 delivery locations, which are relatively small. Therefore, the delivery plan generation unit 12 can use the aforementioned spatio-temporal network model to generate a delivery plan in a practical time. [0078] Next, the outline of the first region division processing will be explained using FIG. 11. First, the first area dividing unit 14 divides the delivery sites w1 to w17 into demand points, supply points, and delivery points that are neither demand points nor supply points. Next, the first area dividing unit 14 attaches a one-to-one correspondence between necessary points and supply points. At this time, the first area dividing unit 14 associates those whose distance (movement time) between the required point and the supply point is close to each other. For example, at the delivery site w4 in FIG. 11, the delivery d1 is 1 more than 1, and at the delivery site w5, the delivery d1 is less than 1. At the delivery site w6, the delivery d1 is more than 1, and at the delivery site w7, the delivery d1 is less than 1. At the delivery site w8, the delivery d2 is more than 1, and at the delivery site w9, the delivery d2 is less than 1. In this case, the first area dividing unit 14 matches the demand with the supply, and the distribution bases close to each other correspond to each other. That is, the first area dividing unit 14 associates the delivery site w4 with the delivery site w5 (as a group 1), associates the delivery site w6 with the delivery site w7 (as a group 2), and associates the delivery site w8 with the delivery site. Correspondence is attached to base w9 (as group 3). The other needs and supply points are also attached and corresponding. [0079] Next, the first area dividing unit 14 is a group in which corresponding demand points and supply points are attached, and groups that are close in distance are collected in plural to generate one area. At this time, if there is no other group in the distance, one group can be used as one area. For example, in the above example, the first area dividing unit 14 calculates the distance between the delivery site w4 of the group 1 and the delivery site w6 of the group 2. Then, if the calculated distance is smaller than the predetermined threshold value, group 1 and group 2 are classified into the same area, and if the calculated distance is greater than the predetermined threshold value, group 1 and group 2 are judged as other areas . For the calculation of the distance between groups, for example, in the case of the distance between group 1 and group 2, you can also calculate the distance between the delivery site w4 and the delivery site w7, or calculate the distance between the delivery site w4 and the delivery site w6, and the delivery site w4~ The distances between the delivery sites w7, the delivery sites w5~w6, and the delivery sites w5~w7 are calculated as the average distance between the group 1 and the group 2. Or, the largest (or smallest) distance among the calculated distances can also be used as the distance between group 1 and group 2. In the case of the example in FIG. 11, the first area dividing unit 14 classifies group 2 and group 3 into the same area j2, and classifies group 1 into another area j1, for example. The first area dividing unit 14 performs the same process for other delivery locations, classifies the delivery locations w1 to w17 for each area, and generates the area j1 to the area j3 shown in FIG. 11. [0080] If the area is generated by the first area division process, the delivery plan generation unit 12 may also target each of the area j1 to the area j3, with respect to the supply quantity of the delivery items at each distribution site in each area , The required quantity, etc. as the initial conditions of the distribution problem, the first spatiotemporal network model or the second spatiotemporal network model is used to generate the distribution plan, but even if the calculation using the spatiotemporal network model is not entirely based on the initial state, It is also possible to use the attachment correspondence between the necessary points and the supply points in the first area division processing as branch information. That is, the delivery plan generation unit 12 generates branch information from the attached corresponding demand point to the supply point, and by connecting these, a delivery plan that meets all needs can be generated. However, only the delivery route connecting the closest delivery sites is too limited, and when viewed from the entire delivery plan moving between multiple sites, it may be far from an optimized state. Therefore, the first area dividing unit 14 performs a process of adding a delivery route that may be selected when the delivery plan generating unit 12 generates branch information. Then, the delivery plan generation unit 12 selects an appropriate delivery route from among all the delivery routes that may be additionally selected as candidates for the delivery routes that connect the delivery routes between the closest delivery sites that meet the needs and supplies. To generate branch information, and generate distribution plans for each region. In FIG. 11, the demand point and the supply point are used as the delivery bases for description, but they can also be the stop points with demand and the stop points with supply respectively. [0081] Next, the first area division processing will be described in detail. Fig. 12 is a second diagram illustrating the first region division processing in the first embodiment of the present invention. The processing method (algorithm) of the first region division processing of the present embodiment will be described below. The first area dividing unit 14 performs various processes in the following procedures. 1. First, each stopping point of each delivery location (the first spatiotemporal network model is a delivery location) is divided into supply points and demand points, and a set A of supply points and a set B of demand points are generated. Set V 1 =A∪B. 2. Next, add a new point V 1 s, t, is set to V = V 1 ∪ {s, t}. 3. From A to B, create a delivery route e 11 to e 19 , and for each delivery route, define a cost function c(e) based on the travel time of the route shown by the delivery route. 4. From s to A, from B to t, make a delivery route, and set the cost of this to zero. Let the set of delivery routes from s to A and B to t be E 2 , and the sum set of E 1 and E 2 as E. 5. The two graphs G=(V, E) are constructed by the processing so far. 6. Calculate the maximum matching of the smallest component of G, that is, the path from s to t where the cost becomes the smallest. More specifically, among the routes from s to t, the number of delivery routes from each supply point included in set A is 1, and the number of delivery routes entering each demand point in set B is one. Next, calculate the delivery route (main problem) where the total of the cost of the delivery route connecting s and each supply point of the set A, each demand point of the set B, and t becomes the smallest. According to 4., the cost from s to each supply point of set A, and the cost from each demand point of set B to t is 0. Therefore, the supply point a1 to supply point a3 and the demand point where the cost becomes the smallest can be obtained. b1~ need to attach the corresponding method of point b3. Here, the cost is a function of the movement time, so the closest supply point can be found to match the demand point one-to-one. In the example of FIG. 12, for example, the combination of the delivery route e 11 connecting a1 and b1, the delivery route e 16 connecting a2 and b3, and the delivery route e 18 connecting a3 and b2 are the delivery routes that have the smallest total cost. In the case, a1 and b1, a2 and b3, and a3 and b2 are one-to-one attached correspondence. 7. Secondly, divide all the stops by the area so that the two corresponding stops are included in the same area. At this time, if the movement time between the two corresponding stops is less than a predetermined value, it is classified into one area. In the example in Figure 12, for example, attach the corresponding a1 and b1, a2 and b3, a3 and b2, with a1 and b1 as one area (j4), and a2 and b3, a3 and b2 as 1 Area (j5) classification of each. [0082] 8. Secondly, once again create two graphs made with the procedures from 1. to 5. to produce the relative problem of 6. (relative problems can be generated by mathematical operations). For branch e=(u,v), if the optimal solution X * (e) of the main problem is 1, then the optimal solution of the relative problem is set to Y, which becomes c(e)-Y * (u)-Y * When (v)=0 (complementary slackness), c^(e)=c(e)-Y * (u)-Y * (v) is used to define c^(e), based on 0 ≦ c^(e) ≦ ε (ε is a predetermined fixed number) so that c^(e) has a width. This condition is not only when c^(e)=0, c^(e) becomes equal to or less than ε, but also means that it is a candidate for delivery route. That is, even in the case where the main problem is not just the least cost formation, a broad solution can be obtained. With the relaxation of the conditions thus solved, in the example of FIG. 12, for example, the candidate e 19 of the delivery route connecting a1 and b1 can be retrieved. [0083] The above first region segmentation processing is complete. Once the first area segmentation process is completed, the delivery plan generation unit 12 will generate a delivery plan for each segmented area. At this time, the delivery plan generating unit 12 uses the candidates of the delivery route obtained in the first area division process to generate a delivery plan that meets the conditions. Specifically, in addition to the above-mentioned "1. Flow preservation rules" ~ "6. Restrictions on branches staying in the delivery base", there are also restrictions on the use of the delivery route obtained in the first area division process. Candidates (e 11 , e 16 , e 18 , e 19 ) calculate the set of branch information. [0084] Next, a description will be given of the flow of the process of generating a delivery plan in this embodiment. Fig. 13 is a flowchart showing an example of the process of generating a delivery plan according to the first embodiment of the present invention. First of all, the delivery personnel performing the delivery plan inputs the initial conditions of the delivery plan to the delivery planning device 10. The input/output unit 13 accepts the input and outputs the accepted information to the initial condition setting unit 11. The initial condition setting unit 11 obtains information on the initial condition input by the delivery person (step S11). The initial condition setting unit 11 sets the acquired initial condition information in the initial condition of the delivery plan. The so-called initial conditions are, for example, the supply quantity (excess quantity), required quantity (shortage quantity) of each site, the number of delivery vehicles at the departure site, the number of delivery personnel, the movement time between each site, and delivery Time limit etc. Next, the first segmentation unit 14 performs the first segmentation process by using the supply quantity and the required quantity of the deliveries at each delivery site (or stopping point) included in the information of the initial conditions (step S12). The first area division processing is as described using FIG. 11 and FIG. 12. [0085] Secondly, the delivery plan generation unit 12 generates each area obtained in the first area division processing in a manner that meets the above-mentioned restriction conditions on the spatiotemporal network model illustrated in FIG. 3 or FIG. For branch information, a plurality of sets of branch information that meet the conditions of the delivery period and the like are generated (step S13). Specifically, the delivery plan generation unit 12 adds information on the number of deliveries, the number of delivery vehicles, and the number of delivery personnel for the candidates of the delivery route obtained in the first area division process, and generates information that meets the respective restriction conditions ( "1. Flow preservation rules" ~ "6. Restrictions on the branches staying in the delivery base") branch information. The delivery plan generation unit 12 combines the generated branch information, and generates a plurality of sets of branch information indicating the number of delivery required by each base until the delivery deadline. [0086] Next, the delivery plan generation unit 12 calculates the total cost for each region and each branch information set after generation (step S14). For example, for each delivery vehicle, delivery personnel, and delivery items, the unit cost incurred per unit time will be pre-recorded in the memory unit 15. The delivery plan generation unit 12 multiplies the unit costs of delivery vehicles, delivery personnel, and delivery items. Calculate the cost of each branch (the total cost of delivery vehicles, delivery personnel, and delivery items) based on the time shown in each branch above. The delivery plan generation unit 12 calculates the cost of each branch included in the set of branch information and adds up the costs. The total cost is the cost for the collection of one branch information. The delivery plan generation unit 12 calculates the cost for each group of all branch information for each area. [0087] Next, the delivery plan generation unit 12 compares the calculated cost for each of the calculated sets for each area, and selects the set of branch information with the smallest total cost (step S15). The set of selected branch information represents the movement of delivery items, delivery means, and delivery personnel over time based on the status of the departure site or each delivery site indicated by the initial conditions (Figure 3 and Figure 8). Therefore, as long as delivery is performed based on the collection of branch information, delivery corresponding to the needs of users becomes possible. In other words, this set of branch information is a distribution plan for obtaining one area after division. Once the delivery plan generation unit 12 completes the generation of the delivery plan for each area, the delivery plan for all the delivery bases given as the initial conditions will be generated. [0088] According to the present embodiment, by dividing the entire distribution site for which initial conditions are given into a collection (regions) of distribution sites located at a short distance, the scale of the distribution problem can be reduced, and the distribution in each region can be reduced. The calculation amount of the generation process of the distribution plan. When the area is divided, the candidates for the delivery route are calculated, and the delivery route is selected from the candidates to generate branch information, so the amount of calculation can be further reduced. Thereby, even if the number of delivery items or delivery sites is large and the delivery problem becomes large-scale, the delivery plan can be generated in a practical time (for example, 10 minutes). The delivery plan generation unit 12 can also generate branch information without using the delivery route candidates in step S13 to generate a delivery plan. The cost of the regional distribution plan can also be added to reconstruct the region. For example, there is a pair of areas where the difference in delivery cost is significantly deviated. By expanding the area where the delivery is completed as soon as possible, and reducing the area where the delivery is completed late, the overall cost can be reduced. [0089] <Second Embodiment> Next, with reference to FIGS. 14 to 18, a delivery planning system of another method (second embodiment) for generating a delivery plan in a practical time for large-scale delivery problems will be described. In the first embodiment, a small-scale area is generated by focusing on relatively close demand points and supply points, and gathers a pair of demand points and supply points. In the second embodiment, area division processing is performed centering on the starting base, and a delivery plan is generated in each divided area. The area division processing of the second embodiment is referred to as second area division processing. [0090] FIG. 14 is a functional block diagram showing an example of the delivery planning system of the second embodiment of the present invention. Among the configurations of the second embodiment of the present invention, the functional parts of the configuration delivery planning device 10 of the first embodiment of the present invention are assigned the same reference numerals, and the description of each is omitted. The delivery planning device 10A of the second embodiment is provided with a second area dividing unit 16 instead of the first area dividing unit 14 of the configuration of the first embodiment. The second area dividing unit 16 obtains a plurality of unit route information indicating a predetermined delivery route and the cost when the delivery route is delivered. The predetermined delivery route is a part of the delivery site from the departure site. Return to the original starting point. The second segmentation unit 16 selects a combination of multiple unit route information combinations, the number of delivery sites included in the combination is within a predetermined number, and the cost is the smallest. The set of departure base and delivery base included in the combination of unit route information is the divided area. The second segmentation unit 16 is realized by the CPU included in the delivery planning device 10A reading a program from the storage unit 15 and executing it. [0091] FIG. 15 is a first diagram illustrating a second area division process of the delivery problem in the second embodiment of the present invention. In FIG. 15, the dots represent the delivery location w1 to the delivery location w17, and the four corners represent the departure location c1 to the departure location c3. In this embodiment, first, after enumerating all conceivable needs and supply patterns (patterns), a plurality of simple delivery routes that meet the needs and supplies of each model are created for each starting base. Here, there are a sufficient number of delivery vehicles (truck) or delivery personnel at departure location c1 to departure location c3, delivery locations w1, w3, w6, w8, and w10 are supply points, and delivery locations w2, w5, w7, w9, w11 is a need point. The types of delivery items related to delivery are set to be the same. In this case, the so-called simple delivery route that meets the needs and supplies of each departure location is, for example, starting from the departure location c1 to the delivery location w1 (R1), picking up the delivery at the delivery location w1, and transporting the delivery to the delivery location w2 ( R2). Once the delivery is completed, return to the delivery route (as delivery route 1) of the departure base c1 (R3). Similarly, the route from the departure site c1 to the departure site c1 via the delivery site w3 and the delivery site w5 (as delivery route 2), and the delivery route from the departure site c1 to the departure site c1 via the delivery site w6 and the delivery site w7 (as Delivery route 3), the delivery route from the departure location c2 to the departure location c2 via the delivery location w8 and the delivery location w9 (as the delivery route 4), and from the departure location c3 to the delivery location c3 via the delivery location w10 and the delivery location w11 Delivery route (as delivery route 5), etc. A plurality of these simple paths are created in advance and recorded in the memory unit 15. Although not shown, for example, the following path is also created and recorded in the storage unit 15 in advance. From the departure location c1 to the delivery location of the departure location c1 via the delivery location w1 and the delivery location w5, from the departure location c1 to the delivery location of the departure location c1 via the delivery location w1 and the delivery location w7, from the departure location c1 to the delivery location w3 And delivery location w2 to return to the delivery location of departure location c1, from departure location c1 to delivery location w10 and delivery location w11 to return to the delivery location of departure location c1 delivery path, from departure location c2 via delivery location w1 and delivery location w2 to return to the departure location The delivery route of c2 is a delivery route, etc., from the departure site c3 to the delivery site c3 via the delivery site w1 and the delivery site w2. [0092] The so-called simple delivery route does not simply refer to a delivery route that returns to the original departure point via one supply point and one demand point (the route that patrols two delivery locations and returns to the original departure point) , It can also start from the departure location c1, take 2 deliveries at the delivery location w1, and deliver each delivery to the delivery location w2 and the delivery location w4, and return to the delivery route of the departure location c1 (example of 3 delivery locations) ). Or, starting from the departure location c1, take a delivery item at the delivery location w1 and deliver it to the delivery location w2, then pick up a delivery item at the delivery location w3, and deliver it to the delivery location w4, back to the delivery route of the departure location c1 (An example of 4 delivery locations). [0093] FIG. 16 is a second diagram illustrating the second area division processing of the delivery problem in the second embodiment of the present invention. As shown in FIG. 16, a plurality of simple delivery routes prepared in advance are recorded in the storage unit 15. Attach a correspondence to a simple delivery route, and record the cost of delivery by the delivery route. The cost is, for example, given as a function of movement time. The more information about the simple delivery route and cost recorded in advance, the more accurate (close to the strict optimal solution) delivery plan can be produced. The information including the simple delivery route illustrated in FIG. 16 and the cost corresponding to the delivery route is called unit route information. [0094] FIG. 17 is a third diagram illustrating the second area division processing of the delivery problem in the second embodiment of the present invention. Next, the second area dividing process of the second area dividing unit 16 will be explained using FIGS. 15 to 17. The premise is that the unit path information illustrated in FIG. 16 is recorded in the storage unit 15. The initial conditions related to the delivery requirements (the required number of each delivery site, the number of supplies, the number of delivery vehicles at the departure site, the number of delivery personnel, etc.) will be given. 1. First, the second segmentation unit 16 reads out from the storage unit 15 the unit route information of the delivery requests (needs) that are given the initial conditions, and includes the unit route information that meets a part of the delivery routes. 2. Next, the second area dividing unit 16 combines the delivery routes included in the read unit route information, and creates a route that meets all the delivery requirements as a provisional solution. Here, as a simple delivery route recorded in advance, a method of finding a combination that meets all requirements and the cost is the cheapest. Although a mathematical technique called a set division route can be used, in the case of this technique, Once the number of simple delivery routes is enlarged, the optimization problem of the combination may not be solved in a practical time. For this reason, the second segmentation unit 16 is among the most simple delivery routes memorized by the memory unit 15. It uses a small number of delivery routes to generate initial solutions, and uses a mathematical technique called column generation to add simple delivery routes. To find the most suitable solution for the delivery route. The few simple delivery routes used to generate initial solutions can be selected in any method. In the initial solution generation, a generally provided solver can be used. At this time, a limit is set on the number of delivery sites that one departure site is responsible for (for example, including demand points and supply points, within 6 sites, etc.). Here, the reason for setting a limit (upper limit) on the number of delivery sites is to reduce the amount of calculation and speed up the processing. The number of limited delivery sites is, for example, that calculations can actually be performed, and the number of delivery sites that can be solved within a practical time is set as the upper limit. For the initial solution, for example, the solution shown in FIG. 15 (delivery route 1 to delivery route 5) can be obtained. [0095] 3. Secondly, the second area dividing unit 16 selects one or more delivery routes from the unselected simple delivery routes by using a column generation method (a well-known mathematical technique), and calculates that the replacement has been Cost reduction when a certain simple delivery route is selected and a simple delivery route after selection. The so-called cost reduction is to subtract the original simple delivery route selected before the replacement from the total cost (the cost column in Fig. 16) recorded in the simple delivery route selected as a re-solution by replacement. The total value of the recorded costs. [0096] 4. The second area dividing unit 16 is to update the existing full delivery route with the simple delivery route selected in 3. if the reduced cost after calculation in 3. is a negative value (it is cheaper after replacement) a part of. For example, regarding delivery to the delivery site w7, in the initial solution shown in Fig. 15, although the delivery route 3 will be selected, the second area division unit 16 replaces the delivery route 3 and calculates and replaces it with the departure site c2 via delivery Cost reduction when the base w6 and the delivery base w7 return to the route of the departure base c2 (delivery route 6). According to Figure 16, the cost of delivery route 3 is 2500, and the cost of delivery route 6 is 1500, so the cost reduction is -1000. Therefore, the second area dividing unit 16 updates the delivery route 3 with the delivery route 6. The updated simple delivery route combination is shown in FIG. 17. In the case of calculating cost reduction, the second area dividing unit 16 selects the delivery route so that the number of bases borne by one departure base becomes within the limit. [0097] This example is an example in which a simple delivery route is moved from the scope of responsibility of the departure site c1 to the scope of responsibility of the departure site c3, but the update of the delivery route is not limited to this example. For example, it is also possible to update the combination of delivery route 1 and delivery route 2 in the route undertaken by the departure site c1 to return from the departure site c1 via the delivery site w3 and the delivery site w2 to the departure site c1 in the state shown in Fig. 15 A method of combining the route from the departure location c1 to the departure location c1 via the delivery location w1 and the delivery location w5. Or, it is also possible to update the combination of the delivery route 1 and the delivery route 2 from the departure site c1 to the delivery site w1, the delivery site w2, the delivery site w3, and the delivery site w5 if there are two targeted deliveries at the delivery site w1. The method of the delivery route back to the departure base c1. [0098] 5. The second region dividing unit 16 repeats the processes of 3. to 4., and terminates the second region dividing process at the point in time when the path that can reduce the cost disappears. Once the second area division process is completed, for each departure location, delivery locations within the restriction (for example, 6 locations) will be associated. The set of multiple delivery sites and departure sites (for example, each of the area j6 to the area j8) associated with the associated plural delivery sites is the area obtained by the second area division process. [0099] Next, a description will be given of the flow of the process of generating the delivery plan of this embodiment. Fig. 18 is a flowchart showing an example of the generation process of the delivery plan in the second embodiment of the present invention. The same processing as in FIG. 13 is briefly described. First, the initial condition setting unit 11 obtains information on initial conditions input by a delivery person or the like (step S11). Secondly, the second area division unit 16 uses the information of the initial conditions to perform the second area division by the number of deliveries supplied at each location, the number required, the number of delivery personnel at the departure location, and the type and number of delivery vehicles. Processing (step S121). The second area division processing is as described using FIGS. 15 to 17. [0100] Next, the delivery plan generation unit 12 is for each area acquired in the second area division process (the departure site acquired in the second area division process and the plural delivery sites associated with the departure site) On the spatio-temporal network model illustrated in Fig. 3 or Fig. 8, branch information is generated in a manner that meets the above restrictions, and a plurality of sets of branch information that meet the conditions of the delivery period, etc. are generated (step S131) . In the example of FIG. 17, the delivery plan generation unit 12 generates a set of plural branch information for each of the area j6 to the area j8. The delivery plan generation unit 12 can generate branch information regardless of the simple delivery route and its combination used in the second area division processing, or it can also use the simple delivery route and its combination to generate branch information. [0101] Next, the delivery plan generation unit 12 calculates the total cost for each set of branch information after generation (step S14). In the example of FIG. 17, the delivery plan generation unit 12 calculates the total cost of each generated branch information set for each of the area j6 to the area j8. Next, the delivery plan generating unit 12 compares the calculated total cost, and selects the set of branch information that minimizes the total cost (step S15). In the example of FIG. 17, the delivery plan generation unit 12 selects the set of branch information that minimizes the total cost for each of the area j6 to the area j8. The selected branch information collection represents the distribution plan of each area. In other words, once the distribution plan generation unit 12 has completed the generation of the distribution plan for each area, the distribution plan for the entire distribution base before the division will be generated. [0102] According to this embodiment, it is possible to divide the delivery areas for each departure location by linking all the delivery locations that are required or supplied in the initial condition to the departure locations one by one. By dividing into small-scale distribution areas, the distribution plan generation process is performed for each of these distribution areas, compared to all the necessary distribution sites that are given in the initial conditions as the object to generate the distribution plan Circumstances, can reduce the amount of calculation necessary for the generation of the distribution plan. With this, even if the number of delivery items or delivery sites is large and the delivery problem becomes large-scale, it is possible to draw up a delivery plan in a practical time (for example, 10 minutes). [0103] <Third Embodiment> With reference to FIGS. 19 to 22, a delivery planning system related to another method (third embodiment) for generating a delivery plan in a practical time for large-scale delivery problems will be described. In the first and second embodiments, the distribution problem is spatially divided (area division) based on the needs and supply information of each distribution site included in the initial conditions, thereby dividing the distribution problem into small-scale distribution problems. Solve small-scale distribution problems to achieve high-speed processing. In this third embodiment, the distribution problem is reduced by time division based on the information of the distribution limit time of the distribution contained in the initial conditions, and a distribution plan is generated for each time after division. [0104] FIG. 19 is a functional block diagram showing an example of the delivery planning system of the third embodiment of the present invention. Among the configurations of the third embodiment of the present invention, the functional parts of the configuration and delivery planning device 10B of the first embodiment of the present invention are assigned the same reference numerals, and the description of each is omitted. The delivery planning device 10B of the third embodiment is provided with a time dividing unit 17 instead of the first area dividing unit 14 of the configuration of the first embodiment. The delivery plan device 10B replaces the delivery plan generation unit 12 and includes a delivery plan generation unit 12a. The time dividing unit 17 divides the delivery restriction time given in the initial condition into plural times, thereby dividing the time for each divided delivery problem into small-scale delivery problems. Each time divided by the time dividing unit 17 is called an interval. The delivery plan generation unit 12a is a section formed by the division of the time division unit 17, and solves the delivery problems set with different objective functions or restrictions to generate a delivery plan. For example, the delivery plan generation unit 12a generates the delivery status of the delivery from the first time of each divided section (the progress status of the delivery at the end of the previous section), and the delivery starts in the section. It may be more distributed to the distribution plan such as the distribution point in need. The last section of the section formed by the division of the time division unit 17 is generated. In the last section, the delivery will be delivered to all of the delivery sites that are undelivered among the necessary delivery sites indicated in the initial conditions. The distribution plan. The time division unit 17 and the delivery plan generation unit 12a are realized by the CPU included in the delivery planning device 10B reading programs from the storage unit 15 and executing them. [0105] FIG. 20 is a first diagram illustrating the time division processing of the delivery problem in the third embodiment of the present invention. The left diagram of Fig. 20 shows the initial state of a certain delivery problem. At the delivery site w1, there are more than two deliveries d1, and at the delivery site w2, there are three more deliveries d1. At the delivery site w3, there are less than four deliveries d1, and at the delivery site w4, there are less than one delivery d1. From this initial state, it can be considered that the delivery of the delivery d1 meets the needs of the delivery site w3 and the delivery site w4 within 120 minutes. In the first and second embodiments, the delivery site w1 to the delivery site w4 are divided into regions. In this embodiment, the delivery restriction time given as the initial condition is divided into 120 minutes. Specifically, the time dividing unit 17 divides the delivery limited time of 120 minutes into plural times (sections). For example, the time dividing unit 17 divides the delivery limit time of 120 minutes into two sections of 60 minutes in the first half and 60 minutes in the second half. The length of the time of one section divided by the time dividing unit 17 and the number of sections may be arbitrary. For example, the time dividing unit 17 may divide 120 minutes into 90 minutes and 30 minutes, or into 3 sections every 40 minutes. The time division unit 17 may be the first section after division by a predetermined time shorter than the delivery restriction time given in the initial condition, and the remaining time may be the last section. [0106] The right diagram of FIG. 20 shows that the time dividing unit 17 divides the delivery limit time of 120 minutes into two sections of 60 minutes each. Once the time division unit 17 divides the delivery restriction time, the delivery plan generation unit 12a generates a delivery plan for the first section formed by the division. At this time, the delivery plan generation unit 12a does not aim to meet all the requirements, but generates a delivery plan that meets as many requirements as possible at the end time of one destination section. The so-called goal is to meet all the requirements. In this example, the delivery d1 is delivered in 60 minutes so that the required quantity of the delivery site w3 and the delivery site w4 can be met. The so-called delivery plan that meets as many requirements as possible, for example, 60 minutes after the start of delivery, the number of deliveries d1 at the delivery site w3 is less than 3, and the delivery site w4 has less than 0 deliveries d1. When a delivery plan is generated, and the number of delivery items d1 at the delivery site w3 is less than 2, and the delivery plan where the delivery items d1 at the delivery site w4 is less than 0 is generated, it means the latter delivery plan. [0107] Next, an explanation will be given of the objective function and restriction set for the first interval. The purpose function described next is generally a case where the time division unit 17 divides the delivery restriction time into N pieces, except for the last section (the Nth section), for all sections (along the flow of time, starting from the first section). 1 section to N-first section) the purpose function used. [0108] (Objective function) Minimize the sum of the following (1) to (7). (1) The cost of moving between locations can suppress the cost of moving between locations. (2) At the time when the supply meets the demand, when the supply meets the demand is promoted as quickly as possible, the options for the solution will be reduced (even if the same delivery is possible even at a later time, by only solving the first delivery , Can reduce the options, and reduce the amount of calculation). (3) The number of stop points that meet the needs is prompted to supply from the stop point of the small number, the options for the solution will be reduced (even if the delivery to the stop number of the big number is no problem , By selecting the stop point with the smallest number, you can reduce the number of solution options and reduce the amount of calculation). (4) The number of users (g[w]) who exist in the delivery base and the delivery means in the following section will prompt the last time of the section under calculation so that the next section is not used in Excessive people and delivery methods for the delivery of the deliveries are moved to the departure base. For example, at the last time of the first section, when there is no delivery means that can be moved by mounting a bicycle, in the second section, since the bicycle is not used for delivery, it is moved to the departure base. (5) In the case of meeting the needs of the point of demand and the supply of the supply point, the number of people and delivery methods existing at such delivery locations encourages that no people or delivery methods will remain at the delivery location that meets the needs. In this embodiment, it is related to the time division of each section (although the delivery status of the previous section is inherited), but the delivery plan is calculated independently, so there may be redundant delivery personnel or delivery methods that are not used in the subsequent sections Left in the distribution base. Therefore, (4) and (5) are added to prevent redundant delivery personnel from remaining at the delivery site. (6) The number of deliveries left at the supply point and the number of deliveries that are insufficient at the point of demand. (7) The number of deliveries that are insufficient at the point of demand. Through (6) and (7), the At the last moment, as much as possible to meet the requirements of the supply. The purpose function is expressed as a case where the sum of the functions corresponding to the contents of each item (1) to (7) is minimized, but it is also possible to multiply each item of the purpose function (each purpose function) by an arbitrary coefficient For weighting. [0109] Next, the relevant restrictions are explained. 1. Flow rate preservation rule (I) In the case of sections other than the last section of the divided sections (1) The flow rate of the delivery item at the beginning of the section, the flow rate from the stop point where there is supply becomes 1 , The flow out of the stop where there is no supply becomes 0. At other points, not at the end of the interval, the outgoing flow is equal to the incoming flow. (2) The flow rate of delivery vehicles is the number of delivery vehicles that are present in the base at the start time of the section. At other points, not at the end of the interval, the outgoing flow is equal to the incoming flow. (3) The flow of delivery personnel is at the start time of the section, and the flow out of the road is the number of delivery personnel present in the base. At other points, not at the end of the interval, the outgoing flow is equal to the incoming flow. [0110] (II) The case of the last section among the divided sections (1) When the flow rate of the delivery is at the start time of the section, the flow rate from the stop point where there is supply becomes 1, and there is never supply The flow out of the stop point becomes 0. At the end of the section, the flow rate to the stop where there is a need becomes 1, and the flow rate to the stop where there is no need becomes 0. At other points, the outgoing flow is equal to the incoming flow. (2) The flow rate of delivery vehicles is the number of delivery vehicles that exist at the base at the time of the start of the delivery of the section. At the end of the section, the traffic entering from the road is the number of delivery vehicles needed at the base. At other points, the outgoing flow is equal to the incoming flow. (3) The flow of delivery personnel is the number of delivery personnel present at the base at the time of the start of the delivery of the section, and the flow out of the road. At the end of the section, the traffic entering from the road is the number of delivery personnel who are in need at the base. At other points, the outgoing flow is equal to the incoming flow. [0111] 2. The capacity limit is 1 or less for the delivery items and the movement of people in the delivery site. [0112] 3. Restrictions within the distribution site 1 (E wp is a collection of branches that represent the delivery, delivery means, and people moving from the road to the stop) (1) When the destination of e is the supply point of d, the delivery does not enter the stop from the road. (2) When the destination of e is the required point of d, the flow rate of the delivery and the delivery personnel is the same value (0 or 1). (3) In all cases, the number of bicycles is less than the number of delivery personnel. [0113] 4. Restrictions within the distribution site 2 (E pw is the collection of delivery items, delivery means, and branches that move people from the stop to the road) (1) When the starting point of e is the supply point of d, the flow of the delivery and the delivery personnel is the same value (0 or 1 ). (2) When the starting point of e is the required point of d, the delivery does not go out from the stopping point to the road. (3) In all cases, the number of bicycles is less than the number of delivery personnel. [0114] 5. Restrictions on variables related to the penalty rule at the last time of the interval For any road w, when the number of people or delivery means not used in the next interval is set to g[w], put Enter the restriction of g[w] ≦ f[w]. f[w] is an arbitrary number. By adjusting the value of f[w], it is possible to change the number of persons and delivery methods that can be allowed to remain at the end of the interval. [0115] 6. Restrictions on the branch from the road to the road (1) When the branch from the road to the road is a branch of a car (there is the possibility of a bicycle), it is restricted that there must be a delivery person in the delivery vehicle. The number of people is less than the number of people that can be boarded, and the number of bicycles that cannot be driven is less than the number of cars that can be loaded. (2) When the branch from the road to the road is a bicycle branch and a walking branch, when it is not a bicycle branch (moving by bicycle or on foot), the sum of the number of bicycles and the number of people moving on foot is equal to moving the branch The number of delivery personnel. (3) When the branch from the road to the road is a bicycle branch, but not a bicycle branch, nor a hiking branch (moving by bicycle), the number of bicycles is equal to the number of delivery personnel moving the branch. (4) When the branch from the road to the road is a pedestrian branch, but not a vehicle branch or a bicycle branch (moving on foot), the number of people traveling on foot is equal to the number of delivery personnel moving the branch. [0116] 7. The restrictions on the branches staying in the delivery site are that there must be people in the car, bicycles cannot be parked on the road, and the total volume of the delivery is less than the load of the truck. [0117] The delivery plan generation unit 12a uses the first or second spatiotemporal network to solve the delivery problem (integer plan problem) that is formulated under the objective function or restriction conditions. In actual calculations, cutting can also be added to achieve higher speed. The right diagram of FIG. 20 shows an example of the result of the delivery plan generation unit 12a generating the delivery plan for the first section. According to this plan, 60 minutes after the start of delivery, there is more than one delivery item d1 at delivery site w1 and delivery site w2, and less than two delivery items d1 at delivery site w3. At delivery site w4, the delivery will be known The deficiency of d1 is eliminated. [0118] The advantages of the method of dividing the time in this way can be cited as the second general point. First, by changing the delivery limit time from 120 minutes to 60 minutes, the number of parameters necessary for calculation can be reduced, and the amount of calculation can be reduced. In Figure 20, for the sake of convenience of explanation, an example is given where the number of delivery locations and delivery items are small, but when calculating that more than 20 deliveries are delivered to 20 delivery locations within 120 minutes, for example, it can be seen that The variable is 41,000 and even limited to 61,000. In contrast, when dividing the time in the first half of 60 minutes to produce the above-mentioned delivery plan that meets as many requirements as possible, for example, the number of variables can be reduced to 14,000, and the restricted number can be reduced to about 18,000. By reducing the number of parameters in this way, the amount of calculation can be reduced and the processing speed can be increased. Next, the process of generating a delivery plan for the second half of 60 minutes will be explained. [0119] FIG. 21 is a second diagram illustrating the time division processing of the delivery problem in the third embodiment of the present invention. Fig. 21 shows the process of generating a delivery plan using the second half 60 minutes after the division. The required quantity and supply quantity of each distribution site at the start time of the second half 60 minutes are sorted. At the delivery site w1 and delivery site w2 in the second half of the start time, there is more than one delivery item d1, and at the delivery site w3, there are less than two delivery items d1. The delivery plan generation unit 12a must use the second half 60 minutes to generate a delivery plan that is to deliver the delivery items d1 of the delivery site w1 and the delivery site w2 to meet the needs of the delivery site w3 within the remaining 60 minutes. Regarding delivery location w4, since the shortage of delivery location w4 is resolved in the first half 60 minutes, it is not included in the delivery problem in the second half 60 minutes. [0120] Next, an explanation will be given of the objective function and restriction set for a delivery problem that targets the last section when the delivery restriction time is divided into N pieces. The purpose function and restrictions in this case are the same as those described using the first spatiotemporal network model (Figure 3) and the second spatiotemporal network model (Figure 8). That is, the delivery plan generation unit 12a generates a delivery that minimizes the cost or the movement time required for delivery among the delivery plans that fulfill all the needs until the last time of the last section. plan. [0121] Taking advantage of the processing of the second half of 60 minutes (the last interval), first, similar to the first half of 60 minutes, by setting the delivery limit time to 60 minutes, the number of parameters and the amount of calculation can be reduced. In this way, the processing speed can be increased. Furthermore, by calculating as many delivery plans as possible to meet the requirements in each section up to the last section, the remaining needs will be reduced, so the scale of the delivery problem that must be solved will be reduced. For example, in the examples shown in FIG. 20 and FIG. 21, not only the number of requirements for each need point is reduced, but also meets the needs at the delivery site w4, and therefore the number of delivery sites is reduced. This can reduce the scale of the distribution problem. When the problem of delivering 20 or more deliveries to 20 delivery locations within 120 minutes is divided into 60 minutes, in the second half of 60 minutes, for example, the number of variables can be reduced to 8000 and the limited number can be reduced Reduce to 12,000 levels. In this way, according to the present embodiment, the reduction of the parameters of the division of time and the delivery plan generated in the interval earlier in time can reduce the scale of the delivery problem in the subsequent interval, so that it can be reduced The production process of the distribution plan is speeded up. In the third embodiment, since no area division is performed, compared with the first and second embodiments, the distribution plan can still be generated by ensuring the distribution of the distribution bases, and a more optimized distribution can be generated. plan. [0122] Next, a description will be given of the flow of the process of generating a delivery plan in this embodiment. Fig. 22 is a flowchart showing an example of the process of generating a delivery plan according to the third embodiment of the present invention. The processing similar to that of FIGS. 13 and 18 will be briefly described. First, the delivery person who performs the delivery plan inputs the initial conditions of the delivery plan to the delivery planning device 10B. The input/output unit 13 accepts the input and outputs the accepted information to the initial condition setting unit 11. The initial condition setting unit 11 obtains information on the initial condition input by the delivery person (step S11). Next, the time division unit 17 performs time division processing using the delivery restriction time included in the information of the initial conditions (step S122). The time division processing is as described using FIG. 20. For example, the time dividing unit 17 divides the delivery restriction time of the initial condition into 2 to 3 pieces. Or, when the time (e.g. 75 minutes) that can be estimated at a certain percentage of the time (for example, 75 minutes) that can be estimated at a certain proportion of the time (for example, 75 minutes) that the delivery to the delivery site in need is based on a rule of thumb, etc. For example, 75 minutes) to divide the original delivery restriction time, and use the remaining time as the final section. Alternatively, the time dividing unit 17 may set a time shorter than the delivery restriction time given in the initial condition by a predetermined time as the first section after division, and use the remaining time as the last section. Secondly, the delivery plan generation unit 12a generates a delivery plan in chronological order for each divided section. First, the delivery plan generating unit 12a sets 1 to a counter variable n (step S123). Next, the delivery plan generation unit 12a sets the delivery problem of the n-th section (step S124). For example, in the case of the first section, the delivery plan generation unit 12a calculates the required number of each delivery site, the number of supplies, the number of delivery methods existing at the departure site, and the number of delivery personnel included in the initial condition information , The time in the first interval is used for the delivery restriction time, to meet as many requirements as possible, and to formulate the integer planning problem based on the objective function and restriction conditions for the generation of the cheapest delivery plan. [0123] Secondly, the delivery plan generation unit 12a uses the first or second spatiotemporal network model illustrated in FIG. 3 or FIG. 8 to generate a plurality of models that meet the various restrictions, and as far as possible A collection of branch information that may satisfy the needs of each delivery site (step S132). Next, the delivery plan generation unit 12a calculates the total cost for each set of branch information after generation (step S14). Next, the delivery plan generating unit 12a compares the calculated total cost, and selects the set of branch information that minimizes the total cost (step S15). With this, the delivery plan for the nth (this time is the first) section will be generated. [0124] Next, the delivery plan generation unit 12a determines whether n is equal to N (step S16). When n is equal to N (step S16; Yes), the delivery plan for all the divided sections will be generated, so the delivery plan generation process is terminated. When n is not equal to N (step S16; No), the delivery plan generation unit 12a adds 1 to n (step S17), and repeats the processing from step S124. [0125] Specifically, for example, when the value of n after adding 1 is 2, the delivery plan generation unit 12a performs the results of the delivery plan generated for the first section in step S124. The required number of delivery locations, the number of supplies, the number of delivery methods that exist at the departure location, the number of delivery personnel, and the time of the second section are used in the delivery limit time to finalize the integer planning problem. At this time, if n is equal to N, the delivery plan generation unit 12a formulates the objective function for the generation of the delivery plan that meets all the requirements, and the integer planning problem of the restriction (the same as the first embodiment, the second The same purpose function in the embodiment). When n is not equal to N, the delivery plan generation unit 12a formulates the objective function for the generation of the delivery plan that meets as many requirements as possible, and the integer plan problem of the restriction. The delivery plan generation unit 12a solves the integer plan problem to generate the nth interval delivery plan. The delivery plan generation unit 12a repeats the processing from step S124 to step S17 until a delivery plan is generated (until n is equal to N) for all sections. All the delivery plans generated sequentially for each of the first section to the Nth (last) section are the delivery plans that meet the requirements given in the initial conditions. [0126] In the flowchart of FIG. 22, the delivery restriction time that is initially given in the initial condition is divided into several sections. For example, the delivery planning process may be configured as a second. 1. First, set an interval shorter than the delivery restriction time given in the initial condition (the first delivery restriction time), and generate a delivery plan that can deliver as much as possible within this interval and can be delivered cheaply. 2. Secondly, for the deliveries that cannot be delivered within the first delivery restriction time, a predetermined length interval (second delivery restriction time) that continues with the first delivery restriction time is set to produce as much as possible within the set interval The distribution plan for the delivery of more delivery items. 3. Repeat the steps from 1. to 2. during the period when the total length of the section from the start of delivery does not exceed the delivery limit time. That is, the solution is found while extending the time (time expansion method). The length of the initial section or the extended section can be arbitrarily set according to the delivery situation. 4. When the total time since the start of the delivery exceeds the delivery restriction time, the next interval will be the last, and the occurrence will be in the last interval (the time of completion of the delivery plan just generated before as the start time, and the first delivery restriction The start time of the time is used as a reference, and the time when the delivery restriction time included in the initial conditions has passed is used as the end time.) A delivery plan that meets all needs and minimizes the cost. [0127] When the delivery plan generation process of this embodiment is actually performed, for a delivery problem with 20 delivery locations and 20 deliveries, the quasi-optimal solution (and strict) can be calculated in a practical time (within 10 minutes). For comparison, the difference of the value of the objective function is less than 10%). [0128] The delivery plan generation method of this embodiment can also be used in the following scenes. For example, when it is necessary to deliver all the vehicles to the parking lot where there is a need within 120 minutes, first generate a delivery plan for the first half 60 minutes. Once the delivery plan is completed, the delivery of vehicles will actually start according to the delivery plan. During the delivery of the vehicle, a delivery plan for the remaining 60 minutes is generated. In this way, by generating and processing parallel delivery plans and delivering, for example, at the scene where car sharing services are performed, time can be effectively used. [0129] The processing procedures of the above-mentioned delivery planning devices 10, 10A, and 10B are stored in a computer-readable recording medium in the form of a program, and the program is read and executed by the computer of the delivery planning system. The above processing. The computer-readable recording medium here means magnetic butterfly, optical disk, CD-ROM, DVD-ROM, semiconductor memory, etc. The computer program can also be distributed to the computer through the communication line, and the computer after receiving the distribution can execute the program. [0130] The above-mentioned program can also be used to realize a part of the aforementioned functions. It is also possible to combine the aforementioned functions with programs that have been recorded in the computer system, so-called differential files (differential programs). The delivery planning devices 10, 10A, and 10B can also be constituted by a single computer, or constituted by a plurality of computers that can be communicably connected. [0131] In addition, within the scope not departing from the gist of the present invention, the constituent elements of the above-mentioned embodiments can be appropriately replaced with well-known constituent elements. The technical scope of this invention is not limited to the above-mentioned embodiment, and various changes can be added without departing from the scope of the invention. [0132] A delivery person is an example of a delivery subject, delivery vehicles, trucks, and bicycles are examples of delivery means, and a shared vehicle for car sharing is an example of delivery items. The first area dividing unit 14, the second area dividing unit 16, and the time dividing unit 17 are each an example of a dividing unit. [Industrial Applicability] [0133] According to the above-mentioned distribution planning system, distribution planning method and program, it is possible to formulate a practical time to minimize the cost or moving time relative to large-scale distribution problems The distribution plan of transformation.
