Do core problems hain: DRIVER MATCHING aur PRICING. Matching ke liye drivers ki live location chahiye — geohash ya quadtree se nearby drivers O(log n) mein milte hain, poora list scan karna galat jawab hai. Batao ki drivers har few seconds location update bhejte hain.

Surge pricing PricingStrategy ka best example hai: base fare + distance + time + surge multiplier (demand/supply ratio se). Ride bhi state machine hai: REQUESTED → DRIVER_ASSIGNED → ARRIVED → IN_PROGRESS → COMPLETED. Driver ke accept na karne par next driver ko offer jaana chahiye — timeout ke saath.

interface DriverMatchingStrategy {
  List<Driver> findNearby(Location pickup, double radiusKm);
}

class GeoHashMatchingStrategy implements DriverMatchingStrategy {
  // Drivers geohash bucket mein indexed — poora scan nahi
  public List<Driver> findNearby(Location p, double r) {
    return geoIndex.query(p, r).stream()
        .filter(d -> d.getStatus() == DriverStatus.AVAILABLE)
        .toList();
  }
}

interface PricingStrategy { Money fare(Ride ride); }
// BaseFare + perKm + perMin + surgeMultiplier(demand/supply)