May 19

Questions about anything?

I've caught up with the grading, so check out my comments on your projects. If you're missing assignments, I encourage you to try to fill in the missing pieces - late work does count for some credit.

Common problems I'm seeing in your final project drafts:

• missing or unclear docstrings (What are the inputs & outputs for this function?)
• not enough context for what the code is doing (definition of variables & structures at the top)
• not obvious what is your code vs what is very close to an online source
• not (yet) enough meat

change making

Today we'll look at one last classic problem, making change, using (a) brute force, (b) greedy goal seeking, and (c) dynamic programming.

• coins.py
• coins_output.txt

The (badly named) "dynamic programming" approach is a generalization of "memoization" which is basically the trick of (a) breaking the problem into smaller sub-problems, and (b) storing the answers to those subproblems so that you don't recalculate them. One example: Finding Fibonnici number fib(n) = fib(n-1) + fib(n-2) using that (recursive) formula, but storing values in a dictionary rather than recalculating them.

Explanations :

• wikipedia: change-making problem
• pythonds: dynamic programming - quite readable; I'll review this one in class

lightening review

What have we covered this term?

• analysis of algorithms: O(), P, NP, and all that
• algorithms: graph (Dijkstra's, spanning path, ...), lots of sorting (merge, bogo, quick, bubble, heapsort), tree search
• algorithm methods: brute force, divide & conquer, greedy, dynamic
• data structures: stacks, queues, hash table, B-tree, tree, graphs, indexed array, linked lists, priority queue, heap
• python (some tricks & Jim's style preferences) ; C (pointers and memory allocation)

aside

• portable stand-alone C data structure library | hackernews discussion