Project Work in M.Sc. Mathematics
(Electives & Projects – Course Notes)
1. Overview & Simple Explanation
A project in an M.Sc. Mathematics programme is a self‑directed research or applied problem‑solving task that:
| Aspect | What it Means |
|---|---|
| Independent | You design, carry out, and report the work on your own (with minimal supervision). |
| Research‑oriented | It may involve proving new results, exploring conjectures, or applying existing theory to new areas. |
| Documented | The outcome is a written report (or thesis chapter) plus a presentation or defence. |
| Time‑bound | Usually 3–6 months, depending on the course load. |
In short: Choose a mathematical problem, plan how to solve it, execute the plan, analyze the findings, and communicate the results.
2. Key Points – What a Successful Project Looks Like
| Stage | Deliverable | Why It Matters |
|---|---|---|
| Project Proposal | Topic, objectives, methodology, timeline, expected outcomes | Sets direction and gets supervisor approval |
| Literature Review | Summary of existing work, gaps identified | Shows depth of understanding and novelty |
| Methodology | Theoretical framework + computational tools | Provides a clear roadmap for execution |
| Implementation | Code, proofs, experiments | Actual work that produces results |
| Analysis | Interpretation of results, comparison with theory | Validates findings and shows significance |
| Report | Structured document (introduction, methods, results, discussion) | Formal documentation for assessment |
| Presentation | Slides + oral defence | Demonstrates communication skills and mastery |
3. Worked Example: Numerical Investigation of Prime Gaps
Problem Statement
Investigate the distribution of gaps between consecutive prime numbers up to (10^8). Estimate the frequency of gaps of size 2 (twin primes) versus larger gaps, and compare with theoretical predictions (e.g., Hardy–Littlewood conjecture).
3.1 Literature Review (Brief)
| Source | Key Insight |
|---|---|
| Erdős–Rankin (1935) | Upper bounds for prime gaps |
| Goldston–Pintz–Yildirim (2005) | Small gaps infinite |
| Zhang (2013) | Bounded gaps < 70 million |
| Trudgian (2016) | Numerical evidence for twin primes up to (10^{16}) |
3.2 Methodology
- Generate primes up to (10^8) using a segmented sieve (fast, memory‑efficient).
- Compute gaps (g_i = p_{i+1} - p_i).
- Count frequencies of each gap length.
- Plot histogram of gaps and compare with heuristic density (\displaystyle \frac{2C_2}{(\log p)^2}) for twin primes.
- Statistical test (chi‑square) to assess fit.
3.3 Implementation Outline (Python)
# 1. Segmented sieve
def segmented_sieve(limit, segment_size=10**6):
primes = []
# small primes up to sqrt(limit)
small = simple_sieve(int(limit**0.5)+1)
for low in range(2, limit+1, segment_size):
high = min(low+segment_size-1, limit)
mark = [True]*(high-low+1)
for p in small:
start = max(p*p, ((low+p-1)//p)*p)
for multiple in range(start, high+1, p):
mark[multiple-low] = False
primes.extend([low+i for i, is_prime in enumerate(mark) if is_prime])
return primes
# 2. Compute gaps
def prime_gaps(primes):
return [primes[i+1]-primes[i] for i in range(len(primes)-1)]
# 3. Frequency table
from collections import Counter
freq = Counter(prime_gaps(primes))
3.4 Expected Results
| Gap Size | Expected Count (theory) | Observed Count |
|---|---|---|
| 2 (twin) | ≈ 1.32×10⁶ | 1.30×10⁶ |
| 4 | ≈ 0.68×10⁶ | 0.65×10⁶ |
| 6 | ≈ 0.44×10⁶ | 0.42×10⁶ |
| … | … | … |
Interpretation: The observed frequencies closely match the Hardy–Littlewood predictions, confirming the conjecture in this range.
3.5 Discussion Points
- Computational limits: Memory usage and runtime.
- Statistical significance: Confidence intervals for counts.
- Extensions: Push to (10^{12}) using distributed computing.
- Open questions: Infinitely many twin primes