bioalgo

bioalgo is a lightweight, pure-Python library for nature-inspired, bio-inspired, and metaheuristic optimization algorithms. It includes both discrete combinatorial solvers (N-Queens, Traveling Salesman Problem) and continuous multimodal function optimizers.

---

pypi-AgEIcHlwaS5vcmcCJDRkNTQ5Yzk5LWQzNTgtNGZiZS1hOGEyLWI0MTNhZmEyMDllZQACKlszLCI4NmU3NTdlNy1iYWJiLTQzNTQtYjhlOC04ZTA3NWY0NThiYTkiXQAABiA_nmoBeOPi25RiDzwqaUokVcrjsG3TeUCAsTq1bUDsNQ

Algorithms Included

Algorithm	Type	Description / Operators	Typical Applications
Hill Climbing	Discrete Local Search	Steepest ascent, conflict evaluation, random-restart heuristics	N-Queens, constraint satisfaction
Simulated Annealing (SA)	Metaheuristic	Metropolis acceptance criterion ($P = e^{-\Delta E / T}$), 2-opt swaps, geometric cooling	Traveling Salesman Problem (TSP), combinatorial routing
Particle Swarm Optimization (PSO)	Swarm Intelligence	Velocity/position updates, cognitive ($\text{pBest}$) and social ($\text{gBest}$) components, inertia damping	Continuous multimodal minimization
Differential Evolution (DE)	Evolutionary	Mutation (DE/rand/1, DE/best/1), binomial crossover ($CR$), one-to-one selection	Non-linear parameter estimation, benchmark optimization
Biogeography-Based Optimization (BBO)	Ecological	Species migration ($\lambda, \mu$), Habitat Suitability Index (HSI), elitism, mutation	Continuous global optimization
Firefly Algorithm (FA)	Swarm Intelligence	Light intensity, distance attractiveness ($\beta(r) = \beta_0 e^{-\gamma r^2}$), randomized step damping	Multimodal landscapes with multiple local optima

---

Installation

Install via pip:

pip install bioalgo

Or install in development/editable mode:

git clone https://github.com/example/bioalgo.git
cd bioalgo
pip install -e .

---

Quickstart

1. Particle Swarm Optimization (PSO)

from bioalgo import ParticleSwarmOptimization
from bioalgo.benchmarks import rastrigin

# Optimize the multimodal Rastrigin function in 5 dimensions
pso = ParticleSwarmOptimization(
    cost_func=rastrigin,
    dim=5,
    bounds=(-5.12, 5.12),
    num_particles=40
)

best_vector, best_cost, history = pso.optimize(max_iter=100)
print(f"Optimal Cost: {best_cost:.6e}")
print(f"Optimal Solution: {best_vector}")

2. Differential Evolution (DE)

from bioalgo import DifferentialEvolution
from bioalgo.benchmarks import ackley

de = DifferentialEvolution(
    cost_func=ackley,
    dim=5,
    bounds=(-32.768, 32.768),
    pop_size=40,
    strategy="best/1/bin"
)

best_vector, best_cost, history = de.optimize(max_generations=100)
print(f"Ackley Minimum: {best_cost:.6e}")

3. Solve N-Queens with Hill Climbing

from bioalgo import NQueensProblem, HillClimbing

problem = NQueensProblem(n=8)
solver = HillClimbing(problem)
solution, restarts, steps = solver.search_with_restarts(max_restarts=50)

print(f"Solved in {restarts} restart(s) and {steps} steps:")
print(problem.render_board(solution))

4. Traveling Salesman Problem (TSP) with Simulated Annealing

from bioalgo import TravelingSalesmanProblem, SimulatedAnnealing

problem = TravelingSalesmanProblem()
sa = SimulatedAnnealing(problem, initial_temp=1000.0, cooling_rate=0.995)

best_tour, best_distance, history = sa.optimize(seed=42)
print(f"Best Tour Length: {best_distance:.2f}")
print("Tour Sequence:", " -> ".join(best_tour))

---

Built-in Benchmark Functions

bioalgo.benchmarks includes standard continuous functions:

• sphere(x)
• rastrigin(x)
• ackley(x)
• rosenbrock(x)
• griewank(x)

---

Command Line Interface (CLI)

bioalgo comes with a command-line interface:

# Solve 8-Queens and render the board
bioalgo nqueens -n 8 --render

# Solve Traveling Salesman Problem
bioalgo tsp --temp 1000 --cooling 0.995

# Optimize a continuous benchmark using PSO, DE, BBO, or FA
bioalgo optimize pso --func rastrigin --dim 5 --iters 100
bioalgo optimize de --func ackley --dim 5 --iters 100
bioalgo optimize bbo --func sphere --dim 5 --iters 100
bioalgo optimize fa --func rastrigin --dim 5 --iters 100

---

Building and Publishing to PyPI

1. Build the Distribution Archives

python -m build

This produces .tar.gz (source distribution) and .whl (built wheel) in dist/.

2. Check Package Metadata

twine check dist/*

3. Test on TestPyPI (Recommended first)

twine upload --repository testpypi dist/*

4. Upload to Production PyPI

twine upload dist/*

---

License

This project is licensed under the MIT License - see the LICENSE file for details.