KTDI Wi-Fi Optimization Project

Background

Kolej Tun Dr. Ismail (KTDI), located in Universiti Putra Malaysia, houses hundreds of students across Blok Mawar and Blok Kenanga. These residential blocks face frequent Wi-Fi disruptions, especially during peak usage times like evenings and weekends. Thick cement walls and uncoordinated router placement contribute to poor signal penetration and overlapping channels. Students rely heavily on stable internet connections for online classes, assignments, and communication, making this a critical issue. An optimized network setup is essential to support their academic and digital needs.

Problem Statement

What We Need

Why It's Important

Problem Modelling

Inputs, Outputs & Constraints

• Inputs: Room matrix, available channels
• Outputs: Channel assignments, router placements
• Constraints: No channel repetition for neighbors, all rooms must be covered

Potential Solutions

Algorithm: Greedy Graph Coloring + Welsh-Powell Heuristic was used

Chosen Algorithm

• Graph Coloring: Models interference between adjacent rooms.
• Greedy Strategy: Assigns lowest non-conflicting channel.
• Welsh-Powell: Sorts rooms by degree for optimized coloring.

Flow Diagram

Pseudocode

Flowchart

Algorithm Implementation

Output Produced:

• adjacency_total.csv: Lists adjacency pairs
• channel_assignment.csv: Lists room number and assigned channel

Verification:

• Tested adjacency constraints across rooms and floors
• Confirmed no adjacent rooms use the same channel

Steps Followed:

1. Created adjacency matrices for rooms across blocks and floors
2. Applied Greedy Graph Coloring algorithm to assign channels
3. Marked and validated that no two connected nodes share the same channel

Optimization & Heuristic

Correctness & Complexity Analysis

Results

Reflection & Lessons Learned

• We learned how to apply abstract concepts such as graph coloring to practical engineering problems.
• The team gained experience in collaborative coding, debugging, and performance tuning using Java.
• Working on this project deepened our understanding of algorithm design, especially in real-world applications like network optimization.
• We faced challenges in modeling adjacency and ensuring correctness across multiple scenarios, which taught us the importance of thorough testing.

Conclusion

We successfully addressed the Wi-Fi connectivity issues in Blok Mawar and Kenanga using a graph-based modeling approach.

By incorporating the Welsh-Powell heuristic, we further reduced the number of routers and channels needed.

Our algorithm ensures minimal signal interference by optimally assigning Wi-Fi channels through Greedy Graph Coloring.

The solution improved overall network coverage, reduced congestion, and enhanced digital accessibility for KTDI residents.