🍃 BioPath Optimizer

Ant Colony Optimization for carbon-neutral delivery route planning AlgoFest Hackathon 2026 · AI/ML + Sustainable Technology

---

The Problem

Global logistics accounts for ~8% of worldwide CO₂ emissions. The root cause is inefficient routing — most delivery systems rely on greedy algorithms that leave massive fuel savings untapped.

BioPath Optimizer solves this using a mechanism nature already perfected: ant colonies.

---

The Biomimicry Insight

When ants forage, they deposit pheromone trails. Shorter paths get traversed more often, accumulating stronger trails, attracting more ants. Over generations, the colony converges on the optimal route — no central controller, no global knowledge needed.

This is Ant Colony Optimization (ACO), and it maps directly onto the Travelling Salesman Problem at the heart of every delivery network.

---

Measured Results

Network	ACO Cost	Greedy Cost	Improvement
5-city	70.00	70.00	Matched optimal
10-city	156.34	208.47	25.01% better

A 25% fuel reduction across a mid-sized logistics fleet (100 vehicles, 50,000 km/year each) eliminates approximately 375,000 litres of fuel and ~1,000 tonnes of CO₂ annually.

---

Technical Implementation

Transition probability

P(i→j) = [τ(i,j)^α × η(i,j)^β] / Σ_k [τ(i,k)^α × η(i,k)^β]

τ(i,j) = pheromone intensity on edge (i,j)
η(i,j) = 1/distance(i,j)   heuristic desirability
α = pheromone weight  (default 1.0)
β = heuristic weight  (default 2.5)

Pheromone update

evaporation:    τ(i,j) ← τ(i,j) × (1 - ρ)
reinforcement:  τ(i,j) ← τ(i,j) + Q/L_k   for each ant k using edge (i,j)

ρ = evaporation rate (default 0.4)
Q = deposit constant (100.0)
L_k = total tour cost of ant k

Complexity Analysis

Operation	Time	Space
Single ant tour	O(n²)	O(n)
One generation (m ants)	O(m·n²)	O(n²)
Full run (t iterations)	O(t·m·n²)	O(n²)

With defaults (t=150, m=30): scales to 50+ cities in under 1 second. Brute force is O(n!) — 10 cities = 3,628,800 routes. ACO finds near-optimal in 4,500 targeted evaluations.

Stack

Layer	Technology
Core engine	C++17, zero external dependencies
Selection	Roulette-wheel proportional probability
Visualization	HTML5 Canvas + Chart.js
Demo	Standalone index.html — open in any browser, no install

---

How to Run

Web demo (no install):

Open index.html in any browser → adjust sliders → click Run

C++ engine:

g++ -O2 -std=c++17 -o biopath biopath_optimizer.cpp
./biopath

---

Project Structure

BioPath-Optimizer/
├── biopath_optimizer.cpp   # Core C++ ACO engine
├── index.html              # Interactive web demo
├── results_5city.json      # C++ output
├── results_10city.json     # C++ output
└── README.md

---

Team

Name	Role
Areeba Qammar	C++ engine, ACO algorithm, architecture
Hifza Sultan	Web demo, visualization, presentation

Tracks: Artificial Intelligence & Machine Learning · Sustainable Technology