KishoreMayank / SuperchargerRoutePlanner

Construct the minimum path through a network of supercharging stations

Supercharger Route Planner

Solution Description:

GOAL: Construct the minimum path through the network of supercharging stations

Main Idea:

• Build out a graph starting at the starting node
• Create edge weights based on travel_time + charging_time
• Travese the graph to determine the shortest path

Build an adjacency list:

• For each node, calculate the great circle distance between every other node
• If within the max range, then add node to adjacency list for that specific node

Build the graph based on the starting node:

• Start at the starting node and conduct a BFS through the adjacency list
• Kept track of the seen nodes that get popped off the queue
• Avoided cycles by keeping track of nodes reachable by the parent node
  • Idea is that if the node is reachable by the parent node, there is already a route added to the graph
• Added a distance heuristic to ensure that the node is not deviating too far from the projected path
• Cost (edge weight) was assigned in one of two ways:
  • If next charger charges faster than the source, only add enough charge to reach the next charger
  • If the next charger charges slower than the source, charge fully and then go to that charger
  • Chose this cost algorithm to minimize the number of stops and minmize time spent at each stop

Run Dijkstra's:

• Finally, with a built graph, we can run Dijkstra's to determine the shortest path
• Build the path and output the built path

Results:

• Was able to successfully create a very fast and fairly accurate approximation algorithm
• Used a distance heuristic in the graph building stage to significantly improve the runtime
• The distance heurisitc is an approximate guess of how far out of path the vehicle can deviate
• Consistenly ended up with results are close to the checker_osx

Run Code:

g++ -std=c++11 -O1 main.cpp network.cpp -o candidate_solution