Scheduling Algorithms

Algorithms and techniques for job scheduling in parallel and distributed systems.

When to Use

• Designing scheduling systems for multi-core or distributed environments

• Research on scheduling optimization

• Comparing heuristic vs. metaheuristic approaches

Algorithm Categories

Heuristic Algorithms

Fast, rule-based approaches that provide good (not optimal) solutions.

HEFT (Heterogeneous Earliest Finish Time)

1. Compute upward rank for each task
2. Sort tasks by rank (descending)
3. For each task in order:
   • Assign to processor that minimizes EFT
   • Use insertion-based scheduling

Complexity: O(n² × m) where n=tasks, m=processors

Best for: Heterogeneous systems with known execution times

Min-Min Algorithm

1. For each unmapped task:
   • Find minimum completion time on each machine
2. Select task with overall minimum time
3. Assign to that machine
4. Repeat until all tasks mapped

Complexity: O(n² × m)

Best for: Load balancing in homogeneous systems

Max-Min Algorithm

Same as Min-Min, but select task with MAXIMUM minimum time

Best for: Reducing makespan when long tasks exist

Metaheuristic Algorithms

Stochastic optimization for complex search spaces.

Genetic Algorithm (GA)

Pseudocode

population = initialize_population() for generation in range(max_generations): fitness = evaluate(population) parents = selection(population, fitness) offspring = crossover(parents) offspring = mutate(offspring) population = survivor_selection(population, offspring) return best_solution(population)

Key Components:

• Encoding: Task-to-machine mapping

• Crossover: Order-based, position-based

• Mutation: Swap, insertion, scramble

Variable Neighborhood Search (VNS)

Pseudocode

solution = initial_solution() while not stopping_condition(): for k in range(k_max): solution_prime = shake(solution, k) solution_double_prime = local_search(solution_prime) if improvement(solution_double_prime, solution): solution = solution_double_prime break # Reset to first neighborhood return solution

Neighborhoods:

• N1: Swap two tasks

• N2: Insert task at different position

• N3: Reverse subsequence

Ant Colony Optimization (ACO)

Pseudocode

initialize_pheromones() while not stopping_condition(): for ant in range(num_ants): solution = construct_solution(pheromones, heuristics) solution = local_search(solution) update_pheromones(solutions, evaporation_rate) return best_solution

Pheromone Update:

τ_ij = (1 - ρ) × τ_ij + Σ(1 / L_k) for best solutions

Hybrid Approaches

GA-VNS Hybrid

1. Use GA for global search
2. Apply VNS as local search operator
3. VNS refines offspring before population update

Benefits:

• GA explores search space

• VNS exploits promising regions

ACO-GA Hybrid

1. ACO constructs initial solutions
2. GA operators (crossover, mutation) diversify
3. Pheromone update from best GA solutions

Performance Metrics

Metric Description

Makespan Total completion time (max)

Flow Time Sum of completion times

Load Balance Variance in machine utilization

Speedup Sequential time / Parallel time

Research Directions

• Unrelated Parallel Machines: Different execution times per machine

• Heterogeneous Multi-core: Varying CPU capabilities

• Energy-aware Scheduling: Minimize power consumption

• Dynamic Scheduling: Tasks arrive online

Resources

• Scheduling Theory Handbook

• Topalouglu et al. - "GA-VNS for Parallel Machine Scheduling"