number of customer (according to Huff model) affect the decreasing of stock => done
Add something (a color for instance) to locate the node in graphstream => done
The surface of a building must affect the maximum quantity of a stock => done
Add graphic with number of batch, quantity of goods on road (perhaps in warehouses and shops too) => done
The logistic provider equilibrate stock of warehouse :
A batch does not follow a supply chain like now but go from a point to another and that's it.
When a customer ask for a restock, the goods need to be already in a local warehouse. => done
Make a connection with graphstream to build a graph of the supply chains :
node = warehouse, final dest and provider
edge = this building deliver this other building in the supply chain
=> done
The road network could be send to Graphstream. => done
Improve the behavior of logistic provider :
They choose warehouse (purchasing or renting) according to their customer location.
They can choose :
- the large warehouse according to his size and not necessarily his location in space
- the local warehouse according to his location and not necessarily his size
- the average warehouse according to this formulae :Let ->AB the vector between the local warehouse and the average one.
Let ->CB be the vector between the large warehouse and the average one.
We try to find B which minimize ||(->AB) + (->CB)||.
=> done
Data of the graphic must be saved on a CSV file => done
stockInRoads must be unbundle by type of batch for data on file => done
When a final dest manager change of logistic provider, his associated supply chain must be removed too = done
A stock and a supply chain belong to a FinalDestinationManager. Need to update in consequences the management of warehouse according to this method.
Moreover, when a FinalDest change of LogisticProvider, the LogisticProvider need to update in consequences the usedWarehouses list.
When the LogisticProvider test if an order is needed, he must find the right supply chain
update supply chain network when logistic provider loose a final dest manager
Develop a new "move" skills with these functions :
- The speed and length of Road must be considered to compute shortest path. But The maximum speed on a road must also affect the speed of the batch agent when he moves.
- make a new algo to compute shortest path: need to consider the break bulks time and not just the length or the speed (useful if we had river network or rail network which are more fast in the
transport but not with break bulk).
- The batch let a trace on the path when he moves which allow to analyze the flows on Graphstream.
- Add an easy way to make a graphic with average distance traveled by batch (and by kind of batch).
Nouvelle stratégie de construction de réseaux : pareil que accessibilité mais basé sur valeur de Huff.
revoir système de calcul du taux de décroissance
indice de performance du système
affichage saturation
Ajout d'autres agents Provider + Freight forwarder (agents qui possèdent une préférence (une proba?) de plutôt faire passer la marchandise par tel ou tel Provider)
Size of building affect break bulk time + include more hazard with international travel.
Add a categorization of product with parameters : the management of warehouses (the supply chain) depends on kind of products : for instance,
the more the product is sold rapidly, the more it wil be stock in a local warehouse. From short term => food, fresh food, dry food, textile, non-food product, high-tech appliances, furnitures => to long term
Some product needs to be manage in particular warehouse (foods for instance), and some other could be manage by particular logistic provider
Update : logistic provider manages goods on hinterland, and forwarding agent manages goods on international ways.
So, a final destination manager have at least two contacts : one or more logistic provider, one or more forwarding agent.
The probability to change the logistic provider and the forwarding agent must depends on his performance.
Add maritime terminal : these are the door of the Seine axis.
