HGA-SalesRouteOptimization

Projct Description

This repository contains the source code and experimental results for the article "Advanced Sales Route Optimization through Enhanced Genetic Algorithms and Real-Time Navigation Systems". This work improves the original approach proposed by Zambrano-Vega et al. (2019) by integrating a Hybrid Genetic Algorithm (HGA) combined with Adaptive Mutation, Tabu Search, and Machine Learning techniques to optimize sales routes efficiently.

The proposed method allows real-time route reoptimization based on traffic conditions, weather, and other external factors. Experimental results demonstrate a 20% reduction in total traveled distance and a 15% improvement in delivery time compared to the original model.

Features

• Hybrid Genetic Algorithm (HGA) for optimized sales route planning.
• Adaptive Mutation Operator to maintain diversity and avoid premature convergence.
• Tabu Search to refine solutions and escape local optima.
• LSTM-based Machine Learning Model for travel time prediction considering real-time traffic and weather conditions.
• Integration with Google Maps API for real-time navigation.

Running the Hybrid Genetic Algorithm

1. Preprocess the data:

   python scripts/procesamiento.py

2. Generate the optimization problem:

   python scripts/problem_gen.py

3. Train the LSTM model (Optional, pretrained model available):

   python scripts/traininLSMT.py

4. Run the Hybrid Genetic Algorithm:

   python scripts/Genetic_algorithm.py

📊 Experimental Results

Performance Evaluation

Method	Total Distance (km)	Delivery Time (min)	Computation Time (s)	Diversity Score
Traditional GA	500	480	120	Low
GA + Adaptive Mutation	450 (↓10%)	432 (↓10%)	130	Medium
GA + Adaptive Mutation + Tabu Search	427 (↓15%)	408 (↓15%)	150	High

The Hybrid Genetic Algorithm with Adaptive Mutation and Tabu Search demonstrates a significant improvement in route efficiency.