QF633 C++ for Financial Engineering - Assignment 3: Order Manager
QF633 Assignment 3
In this assignment, we will explore building a simple but very efficient order manager using the <algorithm> library.
An order is minimally represented by the following data structure:
struct Order {
int price ; // the order ’s price , and int type used instead of double type for simplicity
int open_qty ; // the open qty of the order
std :: string order_id ; // unique order identifier to identify the order
bool is_sell ; // order side , true - buy order , false - sell order
};
An order manager is a class which manages all the open orders, i.e. order with open_qty > 0. The order manager should keep the records in a std::vector in
a sorted manner:
- orders with lower prices should be placed before orders with higher prices
- orders with the same price should be sorted according to the order of
- orders with lower prices should be placed before orders with higher prices
insertion, i.e. FIFO order
class OrderManager {
public :
// add a new record to the order manager
void Add ( int price , int qty , const std :: string & order_id , bool is_sell );
// update the open_qty for an existing order record and delete the record if updated_open_qty == 0
void Update ( const std :: string & order_id , int updated_open_qty );
// print all orders following the order specified , i.e.
// - from lower prices to high prices
// - following FIFO if same price
void PrintAllOrders () const ;
private :
std :: vector <Order > orders_ ;
};
With following code snippets, the expected output is inline commented.
OrderManager order_manager ;
order_manager . Add (100 , 2, "o1", true );
order_manager . Add (99 , 1, "o2", true );
order_manager . Add (100 , 3, "o3", true );
order_manager . Add (101 , 4, "o4", true );
order_manager . PrintAllOrders ();
// price =99 open_qty =1 order_id =o2 is_sell =1
// price =100 open_qty =2 order_id =o1 is_sell =1
// price =100 open_qty =3 order_id =o3 is_sell =1
// price =101 open_qty =4 order_id =o4 is_sell =1
// price =98 open_qty =5 order_id =o5 is_sell =0
// price =100 open_qty =2 order_id =o1 is_sell =1
// price =100 open_qty =1 order_id =o3 is_sell =1
// price =100 open_qty =2 order_id =o7 is_sell =1
// price =101 open_qty =4 order_id =o4 is_sell =1
// price =101 open_qty =1 order_id =o6 is_sell =1
Approach 1: Add and then sort
One obvious approach is to simply append the new record to the std::vector,and then sort the std::vector to the desired order. Can you please given an implementation following this approach?
Hint: please take note the following description from std::sort documentation: Sorts the elements in the range [first, last) in non-descending order. The order of equal elements is not guaranteed to be preserved..
Approach 2: Add while preserving the ordering
The other alternative approach is to insert the new report to the correct position of the std::vector, i.e. firstly find the position to insert the new record while preserving the ordering and secondly to insert the new record to the found location. With this approach, it’s not necessary to call sort separately.
We could linearly scan the std::vector to find the correct position to insert, but a better way is to use binary search. During the lecture, we have introduced std::binary_search which can tell if an element exists in an given range. There are another two important functions in the binary search operations category, namely std::lower_bound and std::upper_bound. Can you use either function to find the correct position to insert the new record for this approach?
The following files have been provided in the zip file:
- order .h: declaration for struct Order
- order . cpp : definition for struct Order
- order_manager .h: declaration for class Order Manager
- order_manager_v1 . cpp : implementation to be provided using approach 1
- order_manager_v2 . cpp : implementation to be provided using approach 2
- main . cpp : main function which gives two simple test cases
- CMakeLists . txt: cmake project file to be provided
You are expected to:
provide an implementation using approach 1 inside order_manager_v1.cpp.
provide an implementation using approach 2 inside order_manager_v2.cpp.
complete the CMakeLists.txt to compile an executable order_manager_v1 for approach 1 and an executable order_manager_v2 for approach 2.
provide a PDF document assignment3.pdf, which describes
– the time complexity for OrderManager::Add, and OrderManager::Update for both approaches.
– rationale of how to make sure the desired ordering is maintained for both approaches
create a zip file as Assignment3_YourName.zip, which contains 1) the pdf report, 2) the completed source codes as well as the CMakeLists.txt, and submit to eLearn.
[Bonus point]: You are encouraged to use functions from <algorithm> rather than writing plain for loop.
