Skip to content

vk1cd314/HeuristicSearch

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

HeuristicSearch

Different Search Strategies implemented with a Heuristic Over Crime Data

Key Components

Adaptive Multivariate Heuristic Function

The heuristic is implemented with crime data from the city vancouver.

h(u, v) = (d_geo(c(u), c(v))/d_max) × w₁ + (|P(c(u), r)| + |P(c(v), r)|) × w₂

Where:

  • h(u, v): Cost estimation between nodes u and v
  • d_geo(c(u), c(v)): Geodesic distance (Haversine) between coordinates
  • d_max: Normalization factor (maximum pairwise distance in dataset)
  • |P(c(u), r)|: Cardinality of points within radius r (density metric)
  • w₁, w₂: Tunable hyperparameters balancing distance vs. density

Hyperparameter Configuration

The current implementation utilizes empirically optimized weights:

w= 10  # Geodesic component weight
w= 90  # Density component weight

This configuration prioritizes node density analysis while maintaining geospatial awareness, creating an effective exploration-exploitation balance.

Performance Benchmarking

Comparative analysis of search algorithms under identical environmental conditions:

Algorithm Time Taken (s) Memory Used (MB)
Greedy Best-First Search 2.2625 5317.78
A* Search 1.5606 5317.78
Weighted A* Search 2.5806 5317.78

A* demonstrates Time efficiency, although all of them tend to use around the same amount of memory, this can be due to implementation.

Heatmap of Potiential Paths

Optimized Trajectory Heatmap

About

Different Search Strategies implemented with a Heuristic Over Crime Data

Resources

License

Stars

0 stars

Watchers

1 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors