JP4234796B2 - Product loading planning apparatus and product loading planning method - Google Patents

Description

【００２１】
ｃｏｉｌ＿ｗｅｉｇｈｔ_ik：パレットｉの製品積載位置ｋ上の製品重量
Ｌ１_ik：パレットｉの製品積載位置ｋ上の長手方向中心から製品重心までの距離
ＣＯＮＳＴＡＮＴ１：長手方向の許容最大モーメント値
Ｐ２_i ≦１のとき ｗ₂ ＝ａ₂
Ｐ２_i ＞１のとき ｗ₂ ＝ｂ₂
ただし、長手方向のモーメント値が長手方向の許容最大モーメント値ＣＯＮＳＴＡＮＴ１を越えた場合のペナルティを考慮してｂ₂ ＞ａ₂ とする。
【００２２】
幅方向バランスの目的関数値は（４）式で表わされる。基本的には、パレット中央に重心がくることを要求される。尚、関数形は、図５長手方向バランスの目的関数値と同様であるため、図示は省略する。
【００２３】
【数２】

【００２４】
ｃｏｉｌ＿ｗｅｉｇｈｔ_ik：パレットｉの製品積載位置ｋ上の製品重量
Ｌ２_ik：パレットｉの製品積載位置ｋ上の幅方向中心から製品径方向中心までの距離
ＣＯＮＳＴＡＮＴ２：幅方向の許容最大モーメント値
Ｐ３_i ≦１のとき ｗ₃ ＝ａ₃
Ｐ３_i ＞１のとき ｗ₃ ＝ｂ₃
ただし、幅方向のモーメント値が幅方向の許容最大モーメント値ＣＯＮＳＴＡＮＴ２を越えた場合のペナルティを考慮してｂ₃ ＞ａ₃ とする。
【００２５】
倉庫分散の目的関数値は（５）式で表わされる。
製品は、通常、工場内の互いに比較的離れた倉庫に分散されて保管されているため、製品のハンドリング時間、コストを考慮し、同一パレット上に積載される製品はできる限り少ない数の倉庫に保管された製品からなることが要求される。
Ｐ４_i ＝ｗａｒｅｈｏｕｓｅ＿ｎｕｍ_i ……（５）
ｗａｒｅｈｏｕｓｅ＿ｎｕｍ_i ：パレットｉの倉庫分散数＋異なる倉庫の製品 数の比から与える点数
棟分散の目的関数値は（６）式で表される。
【００２６】
１つのパレット上の製品は、倉庫の数だけでなく、クレーンスパンごとの棟についても、できるだけ少ない数の棟に保管されていた製品の組合せであることが好ましく、したがって（６）式を最小化することが要求される。
Ｐ５_i ＝ｂｕｉｌｄｉｎｇ＿ｎｕｍ_i ……（６）
ｂｕｉｌｄｉｎｇ＿ｎｕｍ_i ：パレットｉの棟分散＋異なる棟の製品数の比か ら与える点数
図６は、上記（５）式，（６）式の説明図である。
倉庫（棟）分散数が１の場合、目的関数値Ｐ４_i （Ｐ５_i ）は０とする。
倉庫（棟）分散数が１より大きい場合、以下の式より目的関数値Ｐ４_i （Ｐ５_i ）を求める。
Ｐ４_i （Ｐ５_i ）＝｛一定の点数（ここでの例は５点）｝
×｛倉庫（棟）分散数−１｝
＋｛一定の点数｝×｛１／（ＭＡＸ−ＭＩＮ＋１｝
×（ＭＡＸ−同じ倉庫の製品の最大数｝
ＭＡＸ＝｛パレットに積載されている最大製品数｝
−｛倉庫（棟）分散数−１｝
ＭＩＮ＝{ パレットに積載されている最大製品数} ／｛倉庫（棟）分散数}
ここで、「同じ倉庫の製品の最大数」とは、パレットｉに積載する製品が、例 えば倉庫Ａに保管されていた製品１０個、倉庫Ｂに保管されていた製品５個、倉庫Ｃに保管されていた製品５個からなるとき、倉庫Ａの製品数が最大であり、「同じ倉庫の製品の最大数」は１０となる。
【００２７】
また、「パレットに積載されている最大製品数」は、通常は図３に示すように２０であるが、前述したように製品によっては通常の製品２つ分の積載場所が占有され、その場合「パレットに積載されている最大製品数」は２０より少ない数となる。
積み地輸送途上長手方向バランスの目的関数値は（７）式、積み地輸送途上幅方向バランスの目的関数値は（８）式で表される。
【００２８】
Ｐ６_i ：パレットｉの積み地輸送途上長手方向バランスの目的関数値
Ｐ６_1i：パレットが１番目に行く倉庫で積載した後のバランス長手方向の目的関数値
Ｐ６_2i：パレットが２番目に行く倉庫で積載した後のバランス長手方向の目的関数値
……
Ｐ６_(n-1)i：パレットがｎ−１番目に行く倉庫で積載した後のバランス長手方向の目的関数値
Ｐ６_i ＝Ｐ６_1i＋Ｐ６_2i＋……Ｐ６_(n-1)i ……（７）
Ｐ７_i ：パレットｉの積み地輸送途上幅方向バランスの目的関数値
Ｐ７_1i：パレットが１番目に行く倉庫で積載した後のバランス幅方向の目的関数値
Ｐ７_2i：パレットが２番目に行く倉庫で積載した後のバランス幅方向の目的関数値
……
Ｐ７_(n-1)i：パレットがｎ−１番目に行く倉庫で積載した後のバランス幅方向の目的関数値
Ｐ７_i ＝Ｐ７_1i＋Ｐ７_2i＋……Ｐ７_(n-1)i ……（８）
ここでは、パレットｉが倉庫を回る順番も求められる。以下では、先ず、倉庫を回る順番を求めるにあたり必要となる偏荷重の求め方について説明する。
【００２９】
図７は、パレットの一例を示す模式図である。
この図７に示すパレットにはその四隅にＡ，Ｂ，Ｃ，Ｄの４軸が備えられており、一軸あたりの最大能力は５２．５ｔである。偏荷重を求めるあたっては、先ず、重心の位置を求める。重心の位置の求め方については省略する。次に、その重心の位置（ここでは、（−ｐ，ｑ）とする）と総重量Ｍとをもとに以下のようにして偏荷重が求められる。
【００３０】
Ａ軸の偏荷重＝Ｍ×｛Ｌ／２−（−ｐ）｝／Ｌ×｛Ｗ／２＋（ｑ）｝／Ｗ
Ｂ軸の偏荷重＝Ｍ×｛Ｌ／２−（−ｐ）｝／Ｌ×｛Ｗ／２−（ｑ）｝／Ｗ
Ｃ軸の偏荷重＝Ｍ×｛Ｌ／２＋（−ｐ）｝／Ｌ×｛Ｗ／２−（ｑ）｝／Ｗ
Ｄ軸の偏荷重＝Ｍ×｛Ｌ／２＋（−ｐ）｝／Ｌ×｛Ｗ／２＋（ｑ）｝／Ｗ
このようにして求めれた偏荷重をもとに、以下のアルゴリズムに従って、積み地における倉庫を回る順番が決定される。
【００３１】
（１）偏荷重が許容値に入る倉庫へ行く。
（２）偏荷重が許容値に入る倉庫が複数ある場合は、偏荷重が許容値に入る複数の倉庫のうち、パレットを持ち上げる軸の中でも最大荷重をうける軸への荷重値が最も小さい倉庫へ行く。
（３）偏荷重が許容値に入る倉庫がない場合は、回る必要のある全ての倉庫の中から、パレットを持ち上げる軸の中でも最大荷重をうける軸への荷重値が最も小さい倉庫へ行く。
【００３２】
上記を繰り返すことにより、倉庫へ行く順番が決定される。
卸し地輸送途上長手方向バランスの目的関数値は（９）式、卸し地輸送途上幅方向バランスの目的関数値は（１０）式で表される。
Ｐ８_i ：パレットｉの卸し地輸送途上長手方向バランスの目的関数値
Ｐ８_1i：パレットが１番目に行く倉庫で卸した後のバランス長手方向の目的関数値
Ｐ８_2i：パレットが２番目に行く倉庫で卸した後のバランス長手方向の目的関数値
……
Ｐ８_(n-1)i：パレットがｎ−１番目に行く倉庫で卸した後のバランス長手方向の目的関数値
Ｐ８_i ＝Ｐ８_1i＋Ｐ８_2i＋……Ｐ８_(n-1)i ……（９）
Ｐ９_i ：パレットｉの卸し地輸送途上幅方向バランスの目的関数値
Ｐ９_1i：パレットが１番目に行く倉庫で卸した後のバランス幅方向の目的関数値
Ｐ９_2i：パレットが２番目に行く倉庫で卸した後のバランス幅方向の目的関数値
……
Ｐ９_(n-1)i：パレットがｎ−１番目に行く倉庫で卸した後のバランス幅方向の目的関数値
Ｐ９_i ＝Ｐ９_1i＋Ｐ９_2i＋……Ｐ９_(n-1)i ……（１０）
卸し地における倉庫を回る順番の決定は以下の手順で行なう。
【００３３】
（１）偏荷重が許容値に入る倉庫へ行く。
（２）偏荷重が許容値に入る倉庫が複数ある場合は、偏荷重が許容値に入る複数の倉庫のうち、その倉庫で卸した後に、パレットを持ち上げる軸の中でも最大荷重をうける軸への荷重値が最も小さい倉庫へ行く。
（３）偏荷重が許容値に入る倉庫がない場合は、回る必要のある全ての倉庫の中から、その倉庫の中で卸した後に、パレットを持ち上げる軸の中でも最大荷重をうける軸への荷重値が最も小さい倉庫へ行く。
【００３４】
上記を繰り返して、倉庫へ行く順番を決定する。
以上のような、目的関数値を使用し、評価を行なう項目は以下のようにまとめられる。
（Ａ）積載荷重 → 最大化
（Ｂ）積付け製品の重心と輸送装置（パレットやコンテナなど）の重心位置の距離の差 → 最小化
（Ｃ）同一輸送装置（パレットやコンテナなど）上の積付け製品の置場のバラツキ → 最小化
（Ｄ）倉庫間をまたがる輸送における各倉庫での積付け製品の重心と輸送装置（パレットやコンテナなど）の重心位置の距離の差 → 最小化
（パレット上の製品が複数倉庫にまたがるときは、例えばＡ倉庫で製品を積み、次にＢ倉庫へと移動するが、この場合、Ａ，Ｂ倉庫の製品でバランスをとるだけでなく、最初の積み地であるＡ倉庫で積付けた製品のバランスを取ることも必要である。）
また、制約としては一般的には以下のような条件が必要である。
【００３５】
図８はその制約条件を示す模式図である。
（Ｅ）積付け製品の荷重和は輸送装置の許容荷重範囲内であること
（Ｆ）積付けた製品が互いに物理的に配置可能であること（製品間で重なりがないこと）
遺伝的アルゴリズムでは、現実的に問題に対応させた遺伝子列の設計（コード化）が重要である。本実施形態では、１つのパレット上の製品の配列を１つの遺伝子列とみなしてコード化を行なう
図９は、パレット上の製品のコード化の手法を示す模式図である。
【００３６】
パレットに積載する製品の数に相当するとともにそれぞれがそのパレット上の積載位置と対応づけられた配列を用意し、その配列に製品番号を設立する。
次に遺伝的アルゴリズムの全体処理フローを説明する（図２参照）。
まず、初期設定として、各種定数やパラメータの読み込みを行ない（ステップ（ａ））、さらに積載の対象とされる製品のデータを読み込み（ステップ（ｂ））、その後以下のシミュレーションを行なう。
【００３７】
先ずステップ（ｃ）では、全製品をパレットに積載したか否かが判定される。最初は当然積載は終了していないためステップ（ｄ）に進む。
ステップ（ｄ）では、先ず図９に示すようにコード化された遺伝子を多数（例えば１００個）作成する。換言すれば、製品を積載したパレットを多数作成する。この段階では、前述した制約条件を満たす範囲で製品をランダムにパレットに積載してよい。また、最後に最良のパレットを１つのみ選択するのであるから、この段階では異なるパレットには同一の製品が積載されるようコード化されていてよい。
【００３８】
ステップ（ｄ）では、このようにして作成した各コード（各パレット）について、優劣を表わす指標として適応度が求められる。この適応度は（２）式〜（１０）式に示した各目的関数値によって（１）式で表わされる。
次にパレットの適応度を基に、選択淘汰を行なう（ステップ（ｅ））。
図１０は、選択淘汰の概念を示す模式図である。
【００３９】
ここでは適応度の小さい値ほど良いパレットとしているので、
（１）適応度の小さい順にパレットを並べ替える。ここでは、適応度の小さい順にパレット１，パレット２，……，パレット１００とする。
（２）並べ替えた後、最も小さいパレット１をｎ個複製し、次に適応度の小さいパレット２をｎ−１個複製し、順次１つずつ減じ、減じた結果が１になった時、それ以外のパレットは１個ずつ複製する。
【００４０】
ｎの値を変更することによって、選択確率を操作する。こうして、良いパレットを増やしていく。
各々のパレットを複製した後、乱数で交叉を行なうパレットを選択する（ステップ（ｆ））。
図１１は、交叉の概念を示す模式図である。
【００４１】
２つのパレット上の製品を、パレット上のある積載位置を境に交換する。２つのパレット上の製品を相互に交換することで、次世代のパレットが作成される。
交叉後、突然変異を行なう（ステップ（ｇ））。
図１２は、突然変異の概念を示す模式図である。
この突然変異では、パレット上の製品を乱数で選択し、その選択された製品をそのパレットに未積載の財源製品と入れ換える操作を行なう。
【００４２】
突然変異の確率は、以下の２つのパラメータで決定する。
（１）全パレット中で突然変異を起こさせるパレットの数。
（２）一つのパレットの中で、突然変異させる製品の数。
このようにして解の探索空間を広げ、より良いパレットを作成していく。
尚、ここで示した選択淘汰、交叉、突然変異の方法は、ここで述べた方法に特化されるものではなく、他の方法でも目的を達成することは可能である。
【００４３】
上記のようにして解の候補（パレット）を増やした後、必要に応じ再度ステップ（ｄ）に戻り、増やした解の候補（パレット）それぞれについて適応度が計算され、再度、ステップ（ｅ）〜（ｇ）の、選択淘汰、交叉、突然変異の操作が繰り返される。例えばこの操作が１００回繰り返された後、ステップ（ｈ）に進み、それまでに作成された解の候補のうち最良のパレット、すなわち適応度の最も小さいパレットが選択され、ステップ（ｉ）に進んでその選択されたパレットに（シミュレーション的に）積載された製品を財源から外し、ステップ（ｃ）に進んで全製品をパレットに積載したか否か判定し、未積載の製品が残っている時は、その未積載の製品について、上記の処理を繰り返す。このようにして、パレットを１つずつ作成していき、財源の製品が全てパレットに積載されるようにしている。
【００４４】
また、本実施形態では、図２に示すステップ（ｈ）に進むごとにパレットを１つずつ選択したが、このようなパレット作成方法のほか、同時に複数のパレットを作成してもよい。本実施形態では、コード化をパレット１つに対して行っているが、その場合、同時に作成しようとする数のパレット、例えば５つのパレットが連結されたものとして、その５つのパレットに１つのコード化を行なう。その場合、その連結された複数のパレットからなるパレット群が、本発明にいう被積載設備の１単位に対応する。複数のパレット（ここでの例では５つパレット）に１つのコード化を行なうこと以外は上記の処理と同様の処理を行なうことによって、図２に示すステップ（ｈ）に進む毎に同時に５つのパレットを選択することができる。
【００４５】
最後に、製品積載計画装置によって作成された計画結果を示す。
まず、図１４を参照して説明した従来技術と本発明との比較結果を説明する。
図１３は、従来技術と本発明との比較結果を表わした図である。
従来技術では、棟単位で財源を見て計画するので、対象とする範囲は１つの棟のみである。そこで、ここでは、従来技術と本発明の双方について、１つの棟のみを財源と見てシミュレーションを行なった。図１３に示す比較結果を見ると、従来技術に比べ、本発明によって作成された計画の方が、積載率としては７％もパレットの積載率が高い。パレットに対する積載率も換算では、９１％である。また、本発明は全てのパレットにおいて、バランスの制約を満たした計画を作成している。
【００４６】
表１に大規模実験として６倉庫を対象とした本発明を適用した結果を示す。
【００４７】
【表１】

【００４８】
表１に示すように平均積載重量は９４％であり、バランス制約も満足しており、倉庫分散数の平均も２以下であり、十分実用的な計画が作成された。
本発明によって作成された計画は優れた計画であり、その他、製品を追加積付けする処理（既にパレットに製品が積載されているパレットに対して、追加という形で製品を積み付けていく方法や、特急指示の製品をパレットに必ず積み付けるような処理にも、本発明を適用ないし応用することができる。
【００４９】
尚、上記実施形態では、適応度を計算するに当たり、（２）式〜（１０）式の全ての目的関数値を採用したが、これら全てを採用する必要はなく、最小限（２）式〜（４）式を採用すればよく、（２）式〜（４）式に、さらに（５）式〜（６）式、ないし（７）式〜（１０）式を加えると、実行可能な輸送効率のより高い積載計画が立案される可能性が高められる。
【００５０】
【発明の効果】
以上説明したように、本発明によれば、輸送効率の高い積載計画を短時間に立案することができる。
【図面の簡単な説明】
【図１】本発明の製品積載計画装置の一実施形態の概略構成を示すブロック図である。
【図２】遺伝的アルゴリズムの全体フローを示す図である。
【図３】パレット積付け構造の例を示す図である。
【図４】製品積載重量の目的関数値の一例を表した図である。
【図５】長手方向バランスの目的関数値の一例を表わした図である。
【図６】（５）式，（６）式の説明図である。
【図７】パレットの一例を示す模式図である。
【図８】制約条件を示す模式図である。
【図９】パレット上の製品のコード化の手法を示す模式図である。
【図１０】選択淘汰の概念を示す模式図である。
【図１１】交叉の概念を示す模式図である。
【図１２】突然変異の概念を示す模式図である。
【図１３】従来技術と本発明との比較結果を表わした図である。
【図１４】従来の製品積載計画の方法である。
【符号の説明】
１ 財源製品格納装置
２ 製品選択装置
３ 選択済み製品格納装置
４ 積載計画装置
５ 積載計画表示・修正装置[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a product loading planning device for creating a plan for selecting a plurality of products from a product group to be loaded and arranging them on a predetermined loading facility on which a plurality of products are loaded.And product loading planning methodAbout.
[0002]
[Prior art]
As a conventional product loading planning method, product data of a product group to be loaded (hereinafter referred to as “resources”) is represented by a loaded facility (hereinafter referred to as “pallet”) such as a pallet, container, or truck bed. There is a way to assign in order. First, product data of financial resources are sorted in descending order of weight, and products are loaded. At this time, in order to prevent the pallet from being displaced, the pallets are stacked diagonally in descending order of weight (see FIG. 14). In addition, constraints such as pallet size and width and outer diameter of adjacent products are checked. If the constraint is not satisfied, the next candidate is selected from the data sorted in order of weight and loaded.
[0003]
However, with this method, once a product is loaded on a pallet and another product is hit, another effort is required such as loading another product. Further, even if products are loaded diagonally so that the weight on the pallet is not offset, it is not intended to positively secure the weight balance, and a deviation in weight occurs. This method is a trial and error method, and it is difficult to load products with high pallet efficiency and high efficiency.
[0004]
Japanese Patent Laid-Open No. 58-6841 discloses that in order to load various products in order by effectively using the product loadable space determined by the vehicle type, the load is loaded in the product loadable space on the vehicle determined by the vehicle type specification. It is possible to specify a virtual vehicle space having a size corresponding to the type, shape, and quantity, determine the product loading specifications in the virtual vehicle space, and then further add to the space left on the vehicle. Although a method for determining the loading specification of other products by specifying a space of another virtual vehicle is disclosed, this method is a planner assigning by trial and error according to the screen displayed on the computer It does not give a highly optimized loading plan.
[0005]
In JP-A-4-233003, when loading other types of parts on a pallet, in order to optimize the type and quantity of the parts to be loaded and improve the transport efficiency, the attribute data of each part and the pallet for loading the parts A method for storing attribute data and determining the type and quantity of pallets used for transportation and parts loaded on the pallets using attribute data of parts and pallets based on a shipping plan is disclosed. This method also aims to optimize the loading according to a specific search procedure, and cannot always create an optimal loading plan.
[0006]
Furthermore, in Japanese Patent Laid-Open No. 6-305567, a plurality of factors that are expected for the packing form are specified at the time of loading planning, and the number of products loaded on the pallet, products are reduced in order to greatly reduce the conventional trial and error time. As a packing specification, which evaluation factors such as packing density, overall stability, and slip between products are to be emphasized. A method of storing and displaying the package appearance candidates in an order that conforms to the specifications is disclosed. This method tries to solve the multi-objective planning problem efficiently as in the present invention described later. However, in this method, it is necessary for a human to prepare a basic pattern for loading in advance. Since the planned plan is also planned based on a preset pattern, it is not always possible to output an optimal loading plan.
[0007]
[Problems to be solved by the invention]
SUMMARY OF THE INVENTION In view of the above circumstances, the present invention provides a loading plan apparatus capable of creating a loading plan with high transport efficiency when loading a product group to be loaded on a loaded facility.And loading planning methodThe purpose is to provide.
[0008]
[Means for Solving the Problems]
The product loading planning apparatus of the present invention that achieves the above object selects a plurality of products from a product group to be loaded, and the plurality of products are loaded with a predetermined loading facility on which a plurality of products are loaded. In the product loading planning device that creates the plan arranged above,
Creating a plurality of proposals for selecting a plurality of products from the product group and arranging the plurality of products on the loaded facility, and a proposal creation changing means for changing the created proposals based on a genetic algorithm; ,
At least for each plan created or changed by the plan creation change means,
(1) Deviation from the target weight of the total weight of products loaded on the loaded equipment
(2) Deviation of the center of gravity position of the loaded facility with respect to a predetermined first direction on the loaded facility in a state where the product is loaded on the loaded facility from the target center of gravity position;
(3) From the target center-of-gravity position of the center of gravity of the loaded equipment in a state where the product is loaded on the loaded equipment in the predetermined second direction intersecting the first direction on the loaded equipment. Deviation of
Evaluation value calculating means for obtaining an evaluation value based on
And a plan extracting means for extracting a plan close to a desired evaluation value from a plurality of plans created or changed by the plan creating / changing means.
[0009]
Here, in the product loading planning apparatus of the present invention, as the genetic algorithm,
(A) A plan having an evaluation value close to a desired evaluation value is selected by duplicating a plan having an evaluation value close to a desired evaluation value from among a plurality of plans created or changed by the plan creation changing means. Selection 高 め る, which increases the probability of being
(B) For at least two of the plurality of plans created or changed by the plan creation / changing means, a part of the products scheduled to be loaded on the loaded facility is replaced with the two plans. Crossover, and
(C) A part of the products scheduled to be loaded on the loadable facility for at least one of the plural plans created or changed by the plan creation / changing means is used as the loadable facility in the plan. Mutation that replaces a product that will not be loaded
A genetic algorithm including at least one of the operations is preferably employed.
[0010]
Moreover, in the product loading planning apparatus of the present invention, when the product group is distributed and arranged in a plurality of warehouses each having one or more buildings, the evaluation value calculation means is configured to perform the above (1) to ( In addition to 3),
(4) Number of warehouses in which products to be loaded on the loaded equipment are distributed
(5) Number of buildings in which products to be loaded on the loaded equipment are distributed
It is preferable to obtain an evaluation value based on.
[0011]
Furthermore, in the product loading planning apparatus of the present invention, when the product group is distributed and arranged in a plurality of warehouses each having one or more ridges, the evaluation value calculating means includes the above (1) to In addition to (3),
(6) A part of the product in the first direction is loaded on the loading facility in the middle of transporting the loading place where a part of the product loaded on the loading facility is loaded on the loading facility. Deviation of the center of gravity of the loaded equipment in the state of being
(7) A part of the product in the second direction is loaded on the loading facility during the transportation of the loading place where a part of the product loaded on the loading facility is loaded on the loading facility. Deviation of the center of gravity of the loaded equipment in the state of being
(8) A part of the product in the first direction in the course of wholesale land transportation in which a part of the product loaded on the loaded facility is wholesaled from the loaded facility is wholesaled from the loaded facility. Deviation of the center of gravity of the loaded equipment in the state of being
(9) A part of the product in the second direction in the course of transporting the wholesale area in which a part of the product loaded on the loaded facility is wholesaled from the loaded facility is wholesaled from the loaded facility. Deviation of the center of gravity of the loaded equipment in the state of being
It is also a preferable aspect to obtain an evaluation value based on.
[0012]
Although there is an attempt to formulate a truck allocation plan using a genetic algorithm (see Japanese Patent Laid-Open Nos. 7-219920 and 7-230443), an example in which a genetic algorithm is conventionally used for a product loading plan Is not found.
In applying the genetic algorithm to the product loading plan according to the present invention, the evaluation value based on the above (1) to (3), or preferably, in addition to the above (1) to (3), the above (4) to ( Since the evaluation values based on 5) or (6) to (8) above are adopted, a feasible and highly efficient loading plan can be made.
In addition, the product loading planning method of the present invention that achieves the above object selects a plurality of products from a product group to be loaded, and the plurality of products is a predetermined object on which the plurality of products are loaded. In the method of planning the arrangement on the loading equipment,
Using the plan creation change means, a plurality of products are selected from the product group and a plurality of plans for arranging the plurality of products on the loaded facility are created, and the created plan is used as a genetic algorithm. Change based on
For each plan created or changed by the plan creation change means, at least,
(1) Deviation from the target weight of the total weight of products loaded on the loaded equipment
(2) a deviation from a target center of gravity position of the loaded facility in a state where a product is loaded on the loaded facility in a predetermined first direction on the loaded facility; and
(3) A target of the center of gravity position of the loaded equipment in a state in which a product is loaded on the loaded equipment in a predetermined second direction intersecting the first direction on the loaded equipment. An evaluation value based on a deviation from the position of the center of gravity is obtained using an evaluation value calculation means, and a proposal close to a desired evaluation value is selected using a proposal extraction means from a plurality of proposals created or changed by the proposal creation changing means. It is characterized by extracting.
Here, in the product loading planning method of the present invention, the genetic algorithm has an evaluation value close to a desired evaluation value among a plurality of plans created or changed by the plan creation changing means. Evaluation close to the desired evaluation value by duplicating the draft (B) a product that is scheduled to be loaded on the loadable equipment for at least two of a plurality of plans created or changed by the plan creation / changing means. (C) At least one of a plurality of plans created or changed by the plan creation changing means is loaded on the loaded facility. It is preferable that the method includes at least one operation of a mutation that replaces a part of the products scheduled to be performed with a product scheduled to be not loaded on the loadable equipment in the plan.
Further, in the product loading planning method of the present invention, when the product group is distributed in a plurality of warehouses each having one or more buildings, the evaluation value calculating means is used to (1) In addition to (3),
(4) The number of warehouses in which the products loaded on the loaded facility are distributed and
(5) It is preferable to obtain an evaluation value based on the number of buildings in which the products loaded on the loaded facility are distributed,
In the case where the product group is dispersedly arranged in a plurality of warehouses each having one or more buildings, in addition to the (1) to (3), using the evaluation value calculation means,
(6) The part of the product in the first direction in the middle of transporting a part of the product loaded with the part of the product loaded on the part of the loaded equipment is the part of the loaded equipment. Deviation of the center of gravity of the loaded equipment loaded on the center of gravity from the target center of gravity
(7) The part of the product in the second direction in the middle of transporting a part of the product in which a part of the product loaded on the loaded facility is loaded on the loaded facility is the loaded facility. Deviation of the center of gravity of the loaded equipment loaded on the center of gravity from the target center of gravity
(8) The part of the product in the first direction in the course of transporting the wholesale area in which a part of the product loaded on the loaded facility is wholesaled from the loaded facility. Deviation of the center of gravity of the loaded equipment from the target center of gravity difference
(9) The part of the product in the second direction during the transportation of the wholesale area in which a part of the product loaded on the loaded facility is wholesaled from the loaded facility. It is also a preferable mode to obtain an evaluation value based on a deviation from the target center of gravity position of the center of gravity of the loaded equipment in a state of being wholesaled.
[0013]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described.
FIG. 1 is a block diagram showing a schematic configuration of an embodiment of a product loading planning apparatus of the present invention.
The product loading planning apparatus shown in FIG. 1 stores a source product storage apparatus 1 in which a source product is stored, a product selection apparatus 2 that selects a product according to an instructed attribute (transport destination, delivery date, etc.), and stores the selected product. The selected product storage device 3, the load planning device 4 that determines the arrangement of products in consideration of each evaluation item, and the result planned by the load planning device 4, and the load plan display / correction that is corrected by the operator This is a device composed of the device 5.
[0014]
The financial product storage device 1 is a device for storing financial products. The product selection device 2 is a device that selects a necessary product for the loading planning device 4 based on attributes such as a transportation destination and a delivery date from the financial product of the financial product storage device 1. The selected product storage device 3 stores the selected product. The loading planning device 4 is a device for creating a plan for efficiently loading products on a pallet.
[0015]
In the loading plan, it is necessary to solve a large-scale combination optimization problem, and in this product loading planning apparatus, a genetic algorithm is adopted as an algorithm for solving this large-scale combination optimization problem. FIG. 2 shows the overall flow. Detailed description will be given later. A genetic algorithm is basically an optimization algorithm that simulates the evolution process of a living organism. A genetic algorithm is an algorithm that holds a group of candidate solutions and gradually generates good solutions by repeating operations such as “selective selection”, “crossover”, and “mutation” within the group. In the case of this embodiment, products loaded on each loading facility (pallet, container, etc.) are regarded as genes, and operations such as “selective selection”, “crossover”, and “mutation” are repeated for a plurality of genes. By doing so, the optimum loading plan is obtained. That is, an objective function representing the superiority or inferiority of a pallet loaded with products is set, a pallet having a good objective function value is selected, and a pallet having a bad objective function value is selected (selection 淘汰). Moreover, in order to create a better pallet while leaving a good part of the pallet, the products in the pallet are partially exchanged (crossed). In addition, in order to make a new character (product loading pattern) that did not exist on the pallet so far, a part of the product in the pallet is exchanged and an attempt is made to create a better pallet (mutation).
[0016]
The load plan display / correction device 5 is a device that displays the plan result output from the load plan device 4 in a graphic and corrects it by an operator.
FIG. 3 is a diagram illustrating an example of a pallet stacking structure.
In the following example, the problem of loading a product on a pallet having a structure as shown in FIG. That is, products can be stacked on this pallet with a maximum of 20 products in 10 rows × 2 columns. However, depending on the dimensions of the product, the space for two normal products may be occupied, and when such a product is loaded, the maximum number that can be loaded per such product Will be reduced by one.
[0017]
Here, a combination optimization problem for selecting a product to be loaded on a pallet from a product group consisting of a large number of products to be loaded is an evaluation value (here, this is a pallet) obtained by the following equation (1). (Referred to as “fitness”) can be set as a problem to minimize.
Fitness_i = W₁ P1_i + W₂ P2_i + W_Three P3_i + W_Four P4_i
+ W_Five P5_i + W₆ P6_i + W₇ P7_i + W₈ P8_i
+ W₉ P9_i
Fitness_i : Pallet i fitness
P1_i : Objective function value of product loading weight of pallet i
P2_i : Objective function value of the longitudinal balance of pallet i
P3_i : Objective function value of balance in the width direction of pallet i
P4_i : Objective function value of warehouse distribution of pallet i
P5_i : Objective function value of pallet i ridge variance
P6_i : Objective function value of longitudinal balance of pallet i during transportation
P7_i : Objective function value of balance in the width direction during transportation of pallet i
P8_i : Longitudinal balance value of pallet i during transportation of wholesale area
P9_i : Objective function value of the balance in the width direction during transportation of wholesale pallet i
w₁ , W₂ , ……, w₉ : Load factor
Hereinafter, each objective function value of the equation (1) will be described.
[0018]
FIG. 4 is a diagram illustrating an example of the objective function value of the product loading weight.
The objective function value of the product load weight is expressed by equation (2). It is required to load the maximum weight.
Z_i = (MAX_WEIGHT-total_weight_i )
/ MAX_WEIGHT (2)
Z_i P1 when ≧ 0_i = Z_i , W₁ = A₁
Z_i <0, P1_i = | Z_i |, W₁ = B₁
However, considering the penalty of overloading constraint violation b₁ > A₁ And
[0019]
MAX_WEIGHT: Maximum loadable weight of pallet (here 140t)
total_weight_i : Weight of all products loaded on pallet i
FIG. 5 is a diagram illustrating an example of the objective function value of the longitudinal balance.
The objective function value of the longitudinal balance is expressed by equation (3). Basically, it is required to have a center of gravity at the center of the pallet.
[0020]
[Expression 1]

[0021]
coil_weight_ik: Product weight on product loading position k of pallet i
L1_ik: Distance from the center in the longitudinal direction on the product loading position k of pallet i to the product center of gravity
CONSTANT1: Allowable maximum moment in the longitudinal direction
P2_i When ≦ 1, w₂ = A₂
P2_i When> 1 w₂ = B₂
However, in consideration of the penalty when the moment value in the longitudinal direction exceeds the allowable maximum moment value CONSTANT1 in the longitudinal direction, b₂ > A₂ And
[0022]
The objective function value of the width direction balance is expressed by equation (4). Basically, it is required to have a center of gravity at the center of the pallet. Since the function form is the same as the objective function value of the longitudinal balance in FIG. 5, the illustration is omitted.
[0023]
[Expression 2]

[0024]
coil_weight_ik: Product weight on product loading position k of pallet i
L2_ik: Distance from the center in the width direction on the product loading position k of the pallet i to the center in the product radial direction
CONSTANT2: Maximum allowable moment value in the width direction
P3_i When ≦ 1, w_Three = A_Three
P3_i When> 1 w_Three = B_Three
However, considering the penalty when the moment value in the width direction exceeds the allowable maximum moment value CONSTANT2 in the width direction, b_Three > A_Three And
[0025]
The objective function value of warehouse distribution is expressed by equation (5).
Since products are usually distributed and stored in warehouses that are relatively distant from each other in the factory, considering the handling time and cost of products, products loaded on the same pallet should be kept in as few warehouses as possible. It is required to consist of stored products.
P4_i = Warehouse_num_i             ...... (5)
warehouse_num_i : The number of points given from the ratio of the number of warehouses on pallet i + the number of products in different warehouses
The objective function value of the building variance is expressed by equation (6).
[0026]
It is preferable that the product on one pallet is not only the number of warehouses but also the ridges for each crane span, which is a combination of products stored in as few ridges as possible, thus minimizing equation (6) It is required to do.
P5_i = Building_num_i               ...... (6)
building_num_i : Number of points given from the ratio of the distribution of pallet i building + the number of products in different buildings
FIG. 6 is an explanatory diagram of the equations (5) and (6).
When the number of warehouse (building) variance is 1, the objective function value P4_i (P5_i ) Is 0.
When the number of warehouse (building) dispersion is greater than 1, the objective function value P4 is obtained from the following formula._i (P5_i )
P4_i (P5_i ) = {A fixed number of points (in this example, 5 points)}
× {Warehouse (building) dispersion number -1}
+ {Fixed score} × {1 / (MAX−MIN + 1}
X (MAX-maximum number of products in the same warehouse)
MAX = {the maximum number of products loaded on the pallet}
-{Warehouse (building) dispersion number -1}
MIN = {maximum number of products loaded on pallet} / {number of warehouses (buildings) distributed}
Here, “the maximum number of products in the same warehouse” means that the products loaded on the pallet i are, for example, 10 products stored in the warehouse A, 5 products stored in the warehouse B, When there are five stored products, the number of products in the warehouse A is the maximum, and the “maximum number of products in the same warehouse” is 10.
[0027]
In addition, the “maximum number of products loaded on the pallet” is usually 20 as shown in FIG. 3, but depending on the product, the loading space for two normal products is occupied as described above. The “maximum number of products loaded on the pallet” is less than 20.
The objective function value of the balance in the longitudinal direction during the transportation of the loading area is expressed by the equation (7), and the objective function value of the balance in the width direction during the transportation of the loading area is expressed by the expression (8).
[0028]
P6_i : Objective function value of longitudinal balance of pallet i during transportation
P6_1i: Objective function value in the balance longitudinal direction after loading in the warehouse where the pallet goes first
P6_2i: Objective function value in the balance longitudinal direction after loading in the warehouse where the pallet goes second
......
P6_{(n-1) i}: Objective function value in the longitudinal direction of the balance after loading at the warehouse where the pallet goes n-1
P6_i = P6_1i+ P6_2i+ …… P6_{(n-1) i}      ...... (7)
P7_i : Objective function value of balance in the width direction during transportation of pallet i
P7_1i: Objective function value in the balance width direction after loading in the warehouse where the pallet goes first
P7_2i: Objective function value in the balance width direction after loading in the warehouse where the pallet goes second
......
P7_{(n-1) i}: Objective function value in the balance width direction after the pallet is loaded in the n-1st warehouse
P7_i = P7_1i+ P7_2i+ …… P7_{(n-1) i}      ...... (8)
Here, the order in which the pallet i goes around the warehouse is also required. In the following, first, a description will be given of how to obtain an unbalanced load that is necessary for determining the order of going around the warehouse.
[0029]
FIG. 7 is a schematic diagram illustrating an example of a pallet.
The pallet shown in FIG. 7 has four axes A, B, C, and D at its four corners, and the maximum capacity per axis is 52.5 t. In obtaining the offset load, first, the position of the center of gravity is obtained. The method for obtaining the position of the center of gravity is omitted. Next, based on the position of the center of gravity (here, (−p, q)) and the total weight M, an unbalanced load is obtained as follows.
[0030]
Unbalanced load on A-axis = M × {L / 2 − (− p)} / L × {W / 2 + (q)} / W
Unbalanced load of B axis = M × {L / 2 − (− p)} / L × {W / 2− (q)} / W
Unbalanced load of C axis = M × {L / 2 + (− p)} / L × {W / 2− (q)} / W
Unbalanced load of D axis = M × {L / 2 + (− p)} / L × {W / 2 + (q)} / W
Based on the unbalanced load thus obtained, the order of going around the warehouse in the loading area is determined according to the following algorithm.
[0031]
(1) Go to the warehouse where the unbalanced load falls within the allowable range.
(2) If there are multiple warehouses where the unbalanced load falls within the allowable value, out of the multiple warehouses where the unbalanced load falls within the acceptable value, the warehouse with the smallest load value on the shaft that receives the maximum load among the axes that lift the pallet go.
(3) If there is no warehouse where the unbalanced load falls within the allowable value, go to the warehouse where the load value to the axis receiving the maximum load is the smallest among all the warehouses that need to be rotated.
[0032]
By repeating the above, the order of going to the warehouse is determined.
The objective function value of the balance in the longitudinal direction during transportation of the wholesale land is expressed by equation (9), and the objective function value of the balance in the width direction during transportation of the wholesale land is expressed by equation (10).
P8_i : Longitudinal balance value of pallet i during transportation of wholesale area
P8_1i: Objective function value in the balance longitudinal direction after wholesale at the warehouse where the pallet goes first
P8_2i: Objective function value in the longitudinal direction of the balance after wholesale at the warehouse where the pallet goes second
......
P8_{(n-1) i}: Objective function value in the longitudinal direction of balance after pallet wholesale at n-1st warehouse
P8_i = P8_1i+ P8_2i+ …… P8_{(n-1) i}      ...... (9)
P9_i : Objective function value of the balance in the width direction during transportation of wholesale pallet i
P9_1i: Objective function value in the balance width direction after wholesale at the warehouse where the pallet goes first
P9_2i: Objective function value in the balance width direction after wholesale at the warehouse where the pallet goes second
......
P9_{(n-1) i}: Objective function value in the balance width direction after the pallet is wholesaled at the n-1st warehouse
P9_i = P9_1i+ P9_2i+ …… P9_{(n-1) i}      (10)
The order of going around the warehouse in the wholesale area is determined by the following procedure.
[0033]
(1) Go to the warehouse where the unbalanced load falls within the allowable range.
(2) If there are multiple warehouses where the unbalanced load falls within the allowable value, out of the multiple warehouses where the unbalanced load falls within the acceptable value, the shaft that lifts the pallet to the shaft that receives the maximum load will be removed. Go to the warehouse with the smallest load value.
(3) If there is no warehouse where the unbalanced load falls within the allowable range, the load on the shaft that receives the maximum load among the shafts that lift the pallet after all the warehouses that need to be rotated are shipped. Go to the warehouse with the lowest value.
[0034]
Repeat the above to determine the order to go to the warehouse.
The items to be evaluated using the objective function values as described above are summarized as follows.
(A) Load capacity → Maximize
(B) Difference in the distance between the center of gravity of the loaded product and the center of gravity of the transportation device (pallet, container, etc.) → Minimize
(C) Unevenness of storage location for products on the same transport equipment (pallets, containers, etc.) → Minimize
(D) Difference in the distance between the center of gravity of the loaded product at each warehouse and the center of gravity of the transportation device (pallet, container, etc.) in transportation across warehouses → Minimize
(When the products on the pallet span multiple warehouses, for example, the products are loaded in the A warehouse and then moved to the B warehouse. In this case, not only the products in the A and B warehouses are balanced, but the first (It is also necessary to balance the products loaded in the A warehouse, which is the loading area.)
In general, the following conditions are necessary as constraints.
[0035]
FIG. 8 is a schematic diagram showing the constraint conditions.
(E) The load sum of the product to be loaded must be within the allowable load range of the transport device.
(F) Stacked products can be physically arranged with each other (no overlap between products)
In genetic algorithms, it is important to design (encode) a gene sequence that actually corresponds to the problem. In this embodiment, coding is performed by regarding the arrangement of products on one pallet as one gene string.
FIG. 9 is a schematic diagram showing a method for coding products on a pallet.
[0036]
An array corresponding to the number of products loaded on the pallet and corresponding to the loading position on the pallet is prepared, and a product number is established in the array.
Next, the overall processing flow of the genetic algorithm will be described (see FIG. 2).
First, as an initial setting, various constants and parameters are read (step (a)), data of a product to be loaded is further read (step (b)), and then the following simulation is performed.
[0037]
First, in step (c), it is determined whether or not all products are loaded on the pallet. Since the loading is not finished at first, the process proceeds to step (d).
In step (d), first, a large number (for example, 100) of encoded genes are created as shown in FIG. In other words, a large number of pallets loaded with products are created. At this stage, products may be randomly loaded on the pallet as long as the above-described constraints are satisfied. In addition, since only one best pallet is selected at the end, at this stage, different pallets may be coded so that the same product is loaded.
[0038]
In step (d), for each code (each pallet) created in this way, fitness is obtained as an index representing superiority or inferiority. This fitness is expressed by equation (1) by the objective function values shown in equations (2) to (10).
Next, selection is performed based on the fitness of the pallet (step (e)).
FIG. 10 is a schematic diagram showing the concept of the selection basket.
[0039]
Here, the smaller the fitness value, the better the palette.
(1) Rearrange palettes in ascending order of fitness. Here, pallet 1, pallet 2,..., Pallet 100 are set in ascending order of fitness.
(2) After the rearrangement, the n smallest palettes 1 are duplicated, the next smaller fitness palette 2 is duplicated, and the number is reduced one by one. When the reduced result is 1, Duplicate other pallets one by one.
[0040]
Manipulating the selection probability by changing the value of n. In this way, we will increase the number of good pallets.
After duplicating each pallet, a pallet to be crossed with a random number is selected (step (f)).
FIG. 11 is a schematic diagram showing the concept of crossover.
[0041]
Products on two pallets are exchanged at a certain loading position on the pallet. Next-generation pallets are created by exchanging products on two pallets.
After crossover, mutation is performed (step (g)).
FIG. 12 is a schematic diagram showing the concept of mutation.
In this mutation, a product on the pallet is selected with a random number, and the selected product is replaced with a financial product that is not loaded on the pallet.
[0042]
The probability of mutation is determined by the following two parameters.
(1) Number of palettes that cause mutations in all palettes.
(2) Number of products to be mutated in one pallet.
In this way, the search space for solutions is expanded and a better palette is created.
Note that the selection method, crossover method, and mutation method shown here are not specialized to the method described here, and other methods can achieve the object.
[0043]
After increasing the number of solution candidates (pallets) as described above, the process returns to step (d) again as necessary, and the fitness is calculated for each of the increased solution candidates (pallets). The operations of selection selection, crossover, and mutation in (g) are repeated. For example, after this operation is repeated 100 times, the process proceeds to step (h), and the best palette, that is, the palette with the lowest fitness is selected from the solution candidates created so far, and the process proceeds to step (i). In step (c), the product loaded on the selected pallet (simulated) is removed from the source of funds, and it is determined whether or not all the products are loaded on the pallet. Repeats the above process for the unloaded product. In this way, the pallets are created one by one so that all the financial products are loaded on the pallet.
[0044]
Further, in the present embodiment, one pallet is selected each time the process proceeds to step (h) shown in FIG. 2, but a plurality of pallets may be created simultaneously in addition to such a pallet creation method. In this embodiment, encoding is performed for one pallet. In this case, it is assumed that the number of palettes to be created simultaneously, for example, five palettes are connected, and one code is assigned to the five palettes. To do. In that case, the pallet group which consists of the connected several pallet respond | corresponds to 1 unit of the to-be-loaded equipment said to this invention. By performing the same process as the above process except that one encoding is performed on a plurality of pallets (in this example, five pallets), each time the process proceeds to step (h) shown in FIG. A palette can be selected.
[0045]
Finally, the plan result created by the product loading plan device is shown.
First, a comparison result between the prior art described with reference to FIG. 14 and the present invention will be described.
FIG. 13 shows a comparison result between the prior art and the present invention.
In the prior art, since the financial resources are planned in units of buildings, the target range is only one building. Therefore, here, for both the prior art and the present invention, a simulation was performed by considering only one building as a resource. When the comparison result shown in FIG. 13 is seen, compared with the prior art, the plan created by the present invention has a pallet loading rate as high as 7%. The loading rate for the pallet is 91% in terms of conversion. Further, the present invention creates a plan that satisfies the balance constraint for all pallets.
[0046]
Table 1 shows the results of applying the present invention for six warehouses as a large-scale experiment.
[0047]
[Table 1]

[0048]
As shown in Table 1, the average load weight was 94%, the balance constraint was satisfied, the average number of warehouse dispersion was 2 or less, and a sufficiently practical plan was created.
The plan created by the present invention is an excellent plan. In addition, there is a process for additionally loading products (a method of loading products in the form of addition to a pallet that already has products loaded on the pallet, The present invention can also be applied to or applied to a process in which a product designated as an express instruction is always stacked on a pallet.
[0049]
In the above embodiment, all the objective function values of the equations (2) to (10) are adopted in calculating the fitness. However, it is not necessary to adopt all of these, and a minimum of the equations (2) to Formula (4) should be adopted, and by adding Formulas (5) to (6) or Formulas (7) to (10) to Formulas (2) to (4), transportation that is feasible The possibility of a more efficient loading plan is increased.
[0050]
【The invention's effect】
As described above, according to the present invention, it is possible to make a loading plan with high transportation efficiency in a short time.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a schematic configuration of an embodiment of a product loading planning apparatus of the present invention.
FIG. 2 is a diagram showing an overall flow of a genetic algorithm.
FIG. 3 is a diagram illustrating an example of a pallet stacking structure.
FIG. 4 is a diagram illustrating an example of an objective function value of product loading weight.
FIG. 5 is a diagram illustrating an example of an objective function value for longitudinal balance.
FIG. 6 is an explanatory diagram of equations (5) and (6).
FIG. 7 is a schematic diagram showing an example of a pallet.
FIG. 8 is a schematic diagram showing constraint conditions.
FIG. 9 is a schematic diagram showing a method for coding products on a pallet.
FIG. 10 is a schematic diagram showing a concept of a selection basket.
FIG. 11 is a schematic diagram showing the concept of crossover.
FIG. 12 is a schematic diagram showing the concept of mutation.
FIG. 13 is a diagram showing a comparison result between the prior art and the present invention.
FIG. 14 shows a conventional product loading planning method.
[Explanation of symbols]
1 Financial product storage device
2 Product selection device
3 Selected product storage device
4 Loading planning device
5 Loading plan display / correction device

Claims (8)

A financial product product storage device that stores product attribute information data including the transportation destination and delivery date of the product group to be loaded,
A product selection device that selects a plurality of products according to attributes such as a transportation destination and a delivery date from the financial product storage device;
A selected product storage device storing product attribute information data selected by the product selection device;
A plurality of product attribute information data is read from the selected product storage device and arranged on a predetermined loading facility on which a plurality of products are loaded based on the product attribute information data and preset evaluation items. A loading planning device for planning, and
In the product loading plan device comprising the loading result display / correction device for displaying the loading plan result prepared by the loading planning device and manually correcting the result,
The loading planner products corresponding to the plurality of product attribute information data read from the selected product storage unit, the conjunction corresponds to the number of products loaded onto a loading facility, the product number of the product A plurality of proposals as arrangements associated with the loading positions on the loaded equipment are created, and the product number in the created arrangement is used as a gene, the loading arrangement on the loaded equipment is used as a gene locus, and the arrangement is a gene. A draft creation and change means that performs coding as a sequence and changes based on a genetic algorithm,
For each coded data of each plan created or changed by the plan creation changing means, at least,
(1) Deviation from the target weight of the total weight of products loaded on the loaded equipment (2) Products are loaded on the loaded equipment in a predetermined first direction on the loaded equipment A deviation of the center of gravity position of the loaded equipment in a state of being placed from a target center of gravity position; and (3) the covered position of the loaded equipment in a predetermined second direction intersecting the first direction on the loaded equipment. The purpose of the optimum solution search method using a genetic algorithm is to calculate the deviation from the target center of gravity position of the center of gravity of the loaded equipment when the product is loaded on the loading equipment from the product attribute information data. An evaluation value calculation means for obtaining an evaluation value based on the objective function as a function;
A product loading planning device comprising plan extracting means for extracting a plan close to a desired evaluation value from a plurality of plans created or changed by the plan creation changing means. The genetic algorithm is
(A) A plan having an evaluation value close to a desired evaluation value can be obtained by duplicating a plan having an evaluation value close to a desired evaluation value from among the plurality of plans created or changed by the plan creation changing means. Selection 淘汰, which increases the probability of being selected
(B) For at least two of the plurality of plans created or changed by the plan creation / changing means, a part of the products scheduled to be loaded on the loaded facility is exchanged between these two plans. And (c) a part of products planned to be loaded on the loadable facility for at least one of a plurality of plans created or changed by the plan creation changing means, The product loading planning apparatus according to claim 1, wherein the loading planning apparatus includes at least one operation of a mutation for replacing a product that is not to be loaded on the loaded facility in the plan, and is executed in the loading planning apparatus. In the case where the product group is distributed in a plurality of warehouses each having one or more buildings, the evaluation value calculation means is in addition to (1) to (3),
(4) The objective function is the number of warehouses in which the products loaded on the loaded equipment are distributed and (5) the number of buildings in which the products loaded on the loaded equipment are distributed. The product loading planning apparatus according to claim 1, wherein an evaluation value based on the objective function is obtained. In the case where the product group is distributed in a plurality of warehouses each having one or more buildings, the evaluation value calculation means is in addition to (1) to (3),
(6) The part of the product in the first direction in the middle of transporting a part of the product loaded with the part of the product loaded on the part of the loaded equipment is the part of the loaded equipment. Deviation of the center of gravity of the loaded equipment loaded on the target from the target center of gravity position (7) Transportation of the loading place where a part of the product loaded on the loaded equipment is loaded on the loaded equipment Deviation from the target center-of-gravity position of the center-of-gravity position of the loaded equipment in a state where the part of the product is loaded on the loaded equipment in the second direction (8) On the loaded equipment The product in the state where the part of the product is wholesaled from the loaded facility in the first direction during the transportation of the wholesale area where a part of the product loaded on the product is wholesaled from the loaded facility. Deviation of the center of gravity of the loading equipment from the target center of gravity (9) A state in which the part of the product in the second direction is being wholesaled from the loadable equipment in the second place on the way of wholesale land transportation where a part of the product loaded on the equipment is wholesaled from the loadable equipment. 2. The product loading planning apparatus according to claim 1, wherein an evaluation value based on the objective function is obtained by using a deviation of the gravity center position of the loaded equipment from the target gravity center position as an objective function. A financial product product storage device that stores product attribute information data including the transportation destination and delivery date of the product group to be loaded,
A product selection device that selects a plurality of products according to attributes such as a transportation destination and a delivery date from the financial product storage device;
A selected product storage device storing product attribute information data selected by the product selection device;
A plurality of product attribute information data is read from the selected product storage device and arranged on a predetermined loading facility on which a plurality of products are loaded based on the product attribute information data and preset evaluation items. A loading planning device for planning, and
In a method of planning a product loading plan using a loading plan display / correction device for displaying a loading plan result drafted by the loading planning device and manually correcting the result,
The loading planning device corresponds to the number of products loaded on the loading facility with products corresponding to a plurality of product attribute information data read from the selected product storage device, and the product number of the product is assigned to the loading plan device. A plurality of proposals as arrangements corresponding to the loading positions on the loading facility are created, and the product number in the created arrangement is a gene, the arrangement of the loading on the loaded facility is a gene locus, and the arrangement is a gene string And the drafting and changing process to change based on the genetic algorithm,
For each plan created or changed by the plan creation change process, at least,
(1) Deviation from the target weight of the total weight of products loaded on the loaded equipment (2) Products are loaded on the loaded equipment in a predetermined first direction on the loaded equipment A deviation of the center of gravity position of the loaded equipment in a state of being placed from a target center of gravity position; and (3) the covered position of the loaded equipment in a predetermined second direction intersecting the first direction on the loaded equipment. The purpose of the optimum solution search method using a genetic algorithm is to calculate the deviation from the target center of gravity position of the center of gravity of the loaded equipment when the product is loaded on the loading equipment from the product attribute information data. An evaluation value based on the objective function as an evaluation value calculation step,
A product loading planning method, comprising: executing a plan extracting step of extracting a plan close to a desired evaluation value from a plurality of plans created or changed by the plan creating / changing step. The genetic algorithm is
(A) A plan having an evaluation value close to a desired evaluation value can be obtained by duplicating a plan having an evaluation value close to a desired evaluation value from among the plurality of plans created or changed by the plan creation changing means. Selection 淘汰, which increases the probability of being selected
(B) For at least two of the plurality of plans created or changed by the plan creation / changing means, a part of the products scheduled to be loaded on the loaded facility is exchanged between these two plans. (C) a part of products planned to be loaded on the loadable facility for at least one of a plurality of plans created or changed by the plan creation changing means, 6. The product loading planning method according to claim 5, wherein the loading planning apparatus includes at least one operation of a mutation that replaces a product that is not loaded on the loading facility in the plan, and is executed in the loading planning apparatus. In the case where the product group is dispersedly arranged in a plurality of warehouses each having one or more buildings, in addition to the (1) to (3), using the evaluation value calculation means,
(4) The objective function is the number of warehouses in which the products loaded on the loaded equipment are distributed and (5) the number of buildings in which the products loaded on the loaded equipment are distributed. 7. The product loading planning method according to claim 5, wherein an evaluation value based on the objective function is obtained. In the case where the product group is dispersedly arranged in a plurality of warehouses each having one or more buildings, in addition to the (1) to (3), using the evaluation value calculation means,
(6) The part of the product in the first direction in the middle of transporting a part of the product loaded with the part of the product loaded on the part of the loaded equipment is the part of the loaded equipment. Deviation of the center of gravity of the loaded equipment loaded on the target from the target center of gravity position (7) Transportation of the loading place where a part of the product loaded on the loaded equipment is loaded on the loaded equipment Deviation from the target center-of-gravity position of the center-of-gravity position of the loaded equipment in a state where the part of the product is loaded on the loaded equipment in the second direction (8) On the loaded equipment The product in the state where the part of the product is wholesaled from the loaded facility in the first direction during the transportation of the wholesale area where a part of the product loaded on the product is wholesaled from the loaded facility. Deviation of the center of gravity of the loading equipment from the target center of gravity (9) A state in which the part of the product in the second direction is being wholesaled from the loadable equipment in the second place on the way of wholesale land transportation where a part of the product loaded on the equipment is wholesaled from the loadable equipment. 8. The product loading planning method according to claim 5, wherein an evaluation value based on the objective function is obtained by using a deviation of the gravity center position of the loaded equipment from the target gravity center position as an objective function.

Images