CSCI 370 - Lecture 4: Optimization and Design Patterns in Software Engineering

---

Topics Covered

• Optimization Algorithms (Christofides’ Algorithm & TSP)
• AI in Drug Discovery and Development
• Strategy Pattern (Design Pattern)
• Observer Pattern (Design Pattern)

---

Traveling Salesman Problem (TSP) & Christofides’ Algorithm

The Problem

• The Traveling Salesman Problem (TSP) asks: What is the shortest route to visit a set of cities and return to the starting point?
• It is NP-hard; the number of possible routes grows factorially (e.g., 10 cities = ~3.6 million routes).

Real-World Relevance

• Logistics (e.g., UPS delivery optimization)
• Fiber optic cable layout
• Genome sequencing

Christofides’ Algorithm Overview

Christofides’ algorithm offers an approximate solution with a guaranteed upper bound of 1.5 times the optimal route length.

Steps:

1. Minimum Spanning Tree (MST):

   • Connect all nodes (cities) without creating loops.
   • Use shortest possible total edge weight.

2. Find Odd-Degree Vertices:

   • Identify vertices in the MST with an odd number of connections.

3. Minimum Weight Perfect Matching:

   • Pair all odd-degree vertices with minimum-weight connections.

4. Combine to Form Eulerian Circuit:

   • Result is a multigraph with all even-degree vertices, allowing an Eulerian circuit (visits every edge once).

5. Convert to Hamiltonian Circuit:

   • Remove repeated cities to obtain a path visiting each city exactly once.

Benefits:

• Near-optimal solution
• Efficient compared to brute-force

Example Use Case:

A delivery company must visit 15 addresses. Christofides’ algorithm reduces route length, saving fuel and time.

---

AI in Drug Discovery & Healthcare

AI Applications in Healthcare

• Speeds up drug development
• Improves hospital efficiency
• Enables personalized treatment based on patient data

Machine Learning in COVID-19 Detection

• Radiomics + Deep Learning:
  • Extract features from X-rays/CT scans to classify pneumonia
• CAD Systems:
  • Aid radiologists by highlighting affected areas
• Explainability Tools (e.g., SHAP):
  • Help understand AI decisions

Model Performance

• Example model (Random Forest) achieved ~82% accuracy for detecting COVID-19 from imaging data

AI in Vaccine Development

• Predicts protein structures
• Accelerates analysis of vaccine candidates (e.g., Pfizer during COVID)
• Enables scaling up of vaccine production

Limitations

• AI models require large, high-quality datasets
• Poor data = inaccurate predictions

Future Outlook

• Projects like Oracle’s “Stargate” invest in predictive healthcare using AI (e.g., early cancer detection)

---

Strategy Pattern (Design Pattern)

Problem with Inheritance:

• Inheritance makes code tightly coupled.
• Adding new behaviors requires modifying many classes.

Strategy Pattern Solution:

Encapsulate interchangeable behaviors and pass them to objects.

Structure:

• Context: Class using a behavior (e.g., Duck)
• Strategy Interface: Common interface for all behaviors (e.g., FlyBehavior)
• Concrete Strategies: Implement different behaviors (e.g., FlyHigh, NoFly)

Benefits:

• Favor composition over inheritance
• Avoid code duplication
• Open for extension, closed for modification (OCP)
• Add new behaviors without changing existing code

Real-World Analogy:

• Payment system: define Payment interface and implement strategies like Cash, CreditCard, PayPal, etc.

---

Observer Pattern (Design Pattern)

Use Case:

Notifying interested parties (observers) when an object (subject) changes.

Example:

• Subject: Stock Market
• Observers: Wall Street, SEC
• Observers want to be notified when a stock is bought or sold.

Initial Problem:

• Code modification was needed every time a new observer was added.

Observer Pattern Structure:

• Subject Interface: Manages observers (add/remove/notify)
• Observer Interface: Contains notify() method
• Concrete Observers: Implement specific behavior on notification

Benefits:

• Decouples subject from observers
• Open for extension, closed for modification
• Easily add/remove observers without touching core logic

Real-World Example:

• GUI Button in C#: adding/removing click event listeners behaves like the observer pattern

---

Key Software Engineering Principle:

Open-Closed Principle (OCP)

Software entities should be open for extension but closed for modification.

• Add functionality via new code
• Avoid changing existing, working code
• Helps prevent bugs and ensures stability

---

Summary

Lecture 5 focused on optimization and design patterns that emphasize clean, maintainable code. Christofides’ algorithm provides a real-world solution to TSP, while Strategy and Observer patterns illustrate how to write flexible, reusable object-oriented code. These patterns support the Open-Closed Principle, ensuring scalable and bug-resistant systems.