Scribd
Upload
50 views

Lecture1 PDF

This document provides an overview of the CSE 202 Algorithms course. It discusses the course logistics, introduces algorithm design and analysis, and outlines different algorithm design paradigms that will be covered. The key points are: the course will teach how to abstract problems, design algorithms for abstracted problems, and analyze algorithms for correctness and efficiency using paradigms like exhaustive search, greedy algorithms, divide-and-conquer, dynamic programming, hill-climbing, and randomized algorithms. Students will complete homework assignments, a midterm, and a final exam to assess their learning.

Uploaded by

ballechase
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
50 views

Lecture1 PDF

This document provides an overview of the CSE 202 Algorithms course. It discusses the course logistics, introduces algorithm design and analysis, and outlines different algorithm design paradigms that will be covered. The key points are: the course will teach how to abstract problems, design algorithms for abstracted problems, and analyze algorithms for correctness and efficiency using paradigms like exhaustive search, greedy algorithms, divide-and-conquer, dynamic programming, hill-climbing, and randomized algorithms. Students will complete homework assignments, a midterm, and a final exam to assess their learning.

Uploaded by

ballechase
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 32

CSE 202: Design and Analysis

of Algorithms

Lecture 1
!
Logistics
• Instructor: Kamalika Chaudhuri (kamalika@cs)
• Office Hours: F 10:30-11:30am, 4110 EBU3B
• TAs: Chicheng Zhang(chz038@eng), Joseph
Geumlek(jgeumlek@cs)
• Office Hours: TBA
• http://cseweb.ucsd.edu/classes/sp16/cse202-a/
Algorithms

Problem Abstraction Algorithm


An Example
What is the shortest route
Problem:
from San Francisco to San Diego ?

Weighted Graph G
Node: City
Abstraction: Edge: Joins two cities
Edge weight: Distance between cities

What is the shortest path between


San Francisco and San Diego in G?
An Example
What is the shortest route
Problem:
from San Francisco to San Diego ?

Abstraction: Shortest path in a graph

Exhaustive Search,
Algorithm:
Dijkstra’s Algorithm, etc.
Algorithm Design Issues

• Correctness
• Efficiency
Algorithm Design Issues:
Correctness

• Recall: Shortest Path problem in a graph


• Does the algorithm always provide a shortest path?
Algorithm Design Issues:
Efficiency
• Recall: Shortest Path problem in a graph
• Algorithm 1: Iterate over all paths from SD to
SF, and find the one with minimum weight.
• Is this a good algorithm?
• No! Can take exponential(v) time
• Dijkstra’s algorithm takes O(e + vlog v) time
Summary
• In this class, you will learn to:
• Create abstractions from problems
• Design algorithms for abstracted problems
!

• Algorithm design issues we will look at:


• Correctness
• Efficiency
Administrivia
Prerequisites
• Mathematical class
• CSE 101 or equivalent
• Calibration Quiz Friday Apr 1 in class
• Will cover Big O notation, Recurrences, Basic
graph search, Dijkstra’s algorithm
• Calibration Quiz will not count towards grade
Textbook
• Algorithm Design,
by Jon Kleinberg, and Eva Tardos
Syllabus
• Chapters1-3: Read on your own (tested in
calibration quiz)
• Chapter 4: Greedy
• Chapter 6: Dynamic Programming
• Chapter 7: Network Flows
• Chapter 13: Randomized Algorithms
Assessment

• Homeworks (25%)
• Midterm (in-class) (35%)
• Final (in class) (35%)
• Class Participation (5%)
• Calibration quiz (0%)
Midterm and Final
• Midterm - in-class on Wed May 11
• Final - Jun 9 at 11:30am-1pm
!

• If you cannot make these dates, please do


not take the class. We will only make
exceptions for medical emergencies.
Homework Policy

• Homeworks are due in class at the


beginning of lecture
• No late homeworks will be accepted!
Collaboration Policy

• Homeworks should be done in groups of


one or two or three
• Email me the name of your group members
by Fri Apr 8
• If you need a group member, please post on
Piazza group for the class
Collaboration Policy

• You must not look for solutions to


homework problems on the internet
• If you happen to see a homework solution
somewhere, please acknowledge the source!
Collaboration Policy

• If someone (except me or TA) helps you with


your homework, acknowledge the source!
• Remember: Academic dishonesty is taken
very seriously!
• More details on this on the class website
Class Participation

• Participation in In-class Problem Sessions


• Participation on Piazza
• Top 10 answerers on Piazza will get 5%
Grading Policy

Any
! algorithm you present must:

• Be clear and unambiguous (but no code)


• Have a proof of correctness
• Have a running time analysis
• Be relatively efficient
Grading Policy: Clarity
!

• Proofs will be graded by correctness and clarity


• Your arguments should be clear and precise
• More partial credit for clearly written partial solution than
attempt to cover up holes in your argument

• Sample solutions on class website


• We’ll look at some examples in class
Algorithm Design Paradigms
Paradigm 1: Exhaustive Search

• Usually computationally inefficient


• May be useful for some small problems
Paradigm 2: Greedy Algorithms

• Build the solution piece by piece


• Examples: Dijkstra’s algorithm, Kruskal
• Does not always work correctly!
Paradigm 3: Divide and Conquer

• Divide the problem into smaller subproblems


• Solve each subproblem
• Combine the results
• Example: Mergesort
Paradigm 4: Dynamic Programming

• Define a proper notion of subtask


• Solve the subtasks by size
• Once smaller subtasks are solved, larger
ones are easy
• Example: computing the edit distance
Paradigm 5: Hill-climbing

• Start with any solution


• Improve the solution by tweaking it a little
• Continue until optimality
• Example: Flows, linear programming
Paradigm 6: Randomized
Algorithms

• Use some randomness in the algorithm to get out


of bad situations
• Usually very simple and efficient
• May not always succeed, but works with high
probability
• Example: min-cut, MAX-SAT
Intractable Problems

• No efficient solution possible, provably!


• A way out is to find a good approximate solution
Summary: Algorithm Design Paradigms

• Exhaustive Search
• Greedy Algorithms: Build a solution incrementally
• Divide and Conquer: Divide into parts, solve each
part, combine results

• Dynamic Programming: Divide into subtasks,


perform subtask by size. Combine smaller subtasks
to larger ones

• Hill-climbing: Start with a solution, improve it


• Randomized: Use some randomness
Announcement
• Email me the name of your group members
by Fri Apr 8
• No lecture on Wed Mar 30
• Calibration Quiz in class on Fri Apr 1
• Covers Big O, recurrences, basic graph
algorithms (BFS/DFS), Dijkstra’s algorithm
• TA OH will be posted on Piazza

You might also like