Research on emergency scheduling based on improved genetic algorithm in harvester failure scenarios

Area of Science:

• Agricultural Engineering
• Operations Research
• Computational Intelligence

Background:

• Harvesting machine failures pose significant risks to crop yields and economic stability.
• Existing scheduling models often struggle with the dynamic and unpredictable nature of machinery breakdowns.

Purpose of the Study:

• To develop an effective emergency scheduling model for agricultural machinery facing random failures.
• To propose a hybrid optimization algorithm that enhances scheduling efficiency and minimizes disruptions.

Main Methods:

• A hybrid optimization algorithm combining a genetic algorithm (GA) and an ant colony algorithm (ACO) was developed.
• The GA's crossover and mutation methods were enhanced, and ACO principles were integrated to prevent local optima.
• Simulations were performed using field data from Deyang, Sichuan Province, with various harvesting machines experiencing random faults.

Main Results:

• The improved hybrid algorithm significantly reduced optimal comprehensive scheduling costs compared to basic GA and ACO.
• Cost reductions ranged from 14.80% to 47.49% across different scenarios.
• The algorithm demonstrated robust global optimization, high stability, and rapid convergence.

Conclusions:

• The proposed hybrid algorithm offers an effective solution for emergency scheduling of agricultural machinery during failures.
• A visual management system was developed to support optimized agricultural machinery scheduling.