Project 2: FEMA needs your help!

Goal: Implement a spatial data structure

Due: 9/28/2009

Suppose FEMA or some other organization is interested in maintaining a data structure to store a collection of cities and towns. As with most collections, insertion of new cities and search for a given city should be fast operations. In addition, FEMA has some other operations they wish to be fast. In particular, if there is a hurricane somewhere in the country, the data structure needs to be able to quickly produce a list of all cities within, say, a ten mile radius of the hurricane. This is called a range query because we are asking for all objects in the collection within a certain distance of a given location. Second, because cities often need to be evacuated for whatever reason, FEMA needs to be able to quickly determine the closest city to a given location. This is called a nearest neighbor query.

Your program should store a collection of cities. Each city object should have the following information:

city name

x coordinate

y coordinate

To make testing easy, use the following class:

class city

{

public:

string name;

double x;

double y;

};

For your convenience, you may assume all coordinates will be from 0 to 100,000.

Your program should support the basic insert routine:

insert( double x, double y, string name)

To insert a new city at location x, y.

Further, you must implement efficient procedures for range queries and nearest neighbor queries:

list range( double x, double y, double radius ) //this should return the list of all cities within radius miles of point (x, y)

city nearestCity( double x, double y ) //this should return the city closest to city point (x, y).

To provide this functionality efficiently, implement a point-quad tree. Your solution to this problem will be graded on correctness (do the queries yield what they are supposed to) and efficiency ( how many cities had to be searched to compute your range query or find the nearest neighbor ).

For extra credit, generalize the concept of the AVL binary search tree and apply it to your quadtree so that you maintain optimal balance for fast operations.

Please contact me early if you have any questions on this assignment.

To allow for efficient grading, your implementation MUST work with the following test code (In addition, your code will be tested with extremely large test cases to measure the efficiency of your implementation):

void

print_citylist (std::list<city> * plist, std::string message)

{

//std::cout << "" << std::endl;

std::cout << "--------- " << message << " BEGIN ---------" << std::endl;

for (std::list<city>::iterator it = plist->begin(); it != plist->end(); it ++) {

std::cout << "City: '" << (*it).name << "'\t(x=" << (*it).x << ", y=" << (*it).y << ")" << std::endl;

}

std::cout << "--------- " << message << " END ---------" << std::endl;

}

int main()

{

pointQuadTree citytree;

citytree.insert ( 100, 100, "Chicago");

citytree.insert ( 25, 125, "McAllen");

citytree.insert ( 50, 150, "Los Angeles");

citytree.insert ( 75, 130, "Detroit");

citytree.insert ( 150, 50, "New York");

citytree.insert ( 10, 40, "Austin");

city closestCity = citytree.nearestCity (63, 84);

cout << "The closest city is " << closestCity.name << endl;

double x = 50;

double y = 95;

double r = 100;

std::list<city> * closeCities = citytree.in_range (x, y, r);

if (NULL == closeCities) {

return;

}

print_citylist (closecities, "city list");

delete closecities;

return 0;

}