DATASETS:
- Freight cost, mode, weight, dates, vendor, manufacture site
- SAfrica, Africa
- Brazil
- Big dataset n comprehensive
- Supply chain based
- Freight data
- Route
- Route
- Europe Dataset
- Route-> separate source and destination nodes
- https://archive-beta.ics.uci.edu/ml/datasets/3d+road+network+north+jutland+denmark
- https://archive-beta.ics.uci.edu/ml/datasets/vehicle+routing+and+scheduling+problems
- https://networkrepository.com/road.php
- https://www.cs.utah.edu/~lifeifei/SpatialDataset.htm
- https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/CUWWYJ
- https://snap.stanford.edu/data/index.html#road
- https://archive-beta.ics.uci.edu/ml/datasets/cargo+2000+freight+tracking+and+tracing
- Routing IN R
- ARCMAP
- Checkout ARCMAP
- GOOGLE OR
- VRP algoss
- Genetic algo(GA).
- Tabu search(TS).
- Simulated Annealing(SA).
- Ant colony system.
- Particle Swarm Optimisation
- RPS - DRP (distribution req planning)
- RPS - MAS
- Dynamic programming approaches
- SVRP
- DQN (Deep Q Network - combination of RL and CNN)
- OCaPi
- Grey Wolf Optimizer (CGWO)
- INPUT: SRC, DEST, Depots (list)
- STANDARD DATA: Google API (place -> lat, long)
- Depot - frequency of item deliveries; traffic
- Cost - freight weight (balancing load) and transportation cost
- Distance - time => cost factor 1
- Delivery urgency - priority based (Time window penalty cost)
- Fuel efficiency, Carbon emissions
- Freshness degradation (for cold logistics alone)
- Road, traffic, weather
- Vehicle conditions
- Driving hours and rest
- (Fleet management)
- Energy use
- Safety and security
- Economic Health
- Ecosystem impacts
- Factors:
-
Depot: Frequencies of item-calculated from dataset using DepotId Used for quick transport of goods Frequent depot-> delivery started as and when the orders are placed Less frequent depot-> delivery process must wait until some amt -> reached.
-
Cost:
-
Distance and time: Optimal path -> less distance and less time.
-
Delivery urgency: If an order needs urgent delivery(Ex: amazon prime subscribers vs normal users), routing must be planned accordingly. cost -> increase
-
Fuel efficiency, carbon emissions: (is there any cnxn bw traffic condition and fuel efficiency?) Suggest route which decrease the carbon emissions. and same with fuel efficiency.
-
Freshness Degradation: When the order is veggies,.. consider the time
-
Road, Traffic and weather: Main problem is we have to get real time data for predicting the current traffics, weather and road condition.
-
Vehicle conditions: Depends on the size of vehicle. if it is a big truck, we can only suggest highways....
-
Driving hours and rest: Consider rest time
-
Safety and security: may be we can consider the safety of the goods ( type of goods) Packaging cost
-
Economic health:
-
Ecosystem impacts: similiar to carbon emissions
-
Capacity of the vehicle
-
IP: Set of orders and also factors
OP: Total cost after considering the factors and Optimal path
(Hierarchy)
-
OUTPUT: Routes + time factor => best k solutions
-
METHODOLOGY: Equation based on all factors - try to optimise the score
-
Challenges
-
IDEA: diverse adaption for any country based on standard data
-
IDEA: SPARK DISTRIBUTED processing
- Finalise mode: Air route, ship route, road route ; or mixed mode
- Finalise scope: India or US or specific country; worldwide, intercontinent
- Identify more factors
- Find equation
-
Route optimise --- > existing route => optimise
-
Graph dataset =>
-
Lat and Longitude
-
Distributed data.... spark
-
Maximise delivery Minimise time Minimise cost