legendtkl Requirements for Outputs
General Guidelines
- PDDL domain files must meet optimization standards.
- Problem files must adhere to the transformations specified.
- Outputs should clearly indicate optimization metrics.
Optimization Behavior
- Optimization processes should complete within a predefined time.
- If no optimization is possible, return an indication of failure.
PDDL Optimization Utilities
1. Load and Transform Domain
load-and-transform-domain(domain_path)
Description:
Loads a PDDL domain file and applies transformations to optimize the state representation.
Parameters:
domain_path(str): Path to the PDDL domain file.
Returns:
optimized_domain: A transformed domain for better efficiency.
Example:
optimized_domain = load_and_transform_domain("domain.pddl")
Notes:
- Utilizes
unified_planning.io.PDDLTransformer. - Throws an error if transformation fails.
2. Optimize Problem
optimize-problem(problem_path, optimized_domain)
Description: Optimizes the specified problem based on the optimized domain loaded previously.
Parameters:
problem_path(str): Path to the PDDL problem file.optimized_domain: The transformed domain.
Returns:
optimized_problem: An optimized problem object ready for planning.
Example:
optimized_problem = optimize_problem("task01.pddl", optimized_domain)
Notes:
- Uses
unified_planning.shortcuts.ProblemOptimizer. - Returns None if no optimization is possible.
3. Save Optimized Problem
save-optimized-problem(optimized_problem, output_path)
Description: Writes the optimized problem to disk in standard PDDL format.
Parameters:
optimized_problem: A PDDL problem that has been optimized.output_path(str): Output file path.
Example:
save_optimized_problem(optimized_problem, "optimized_task01.pddl")
Notes:
- Uses
unified_planning.io.PDDLWriterfor output. - Outputs a text file with optimized specifications.
4. Validate Optimization
validate-optimization(optimized_problem)
Description: Validates that the optimizations applied to the problem are effective and correct.
Parameters:
optimized_problem: The problem that has undergone optimization.
Returns:
- bool: True if optimizations are valid, False otherwise.
Example:
is_valid = validate_optimization(optimized_problem)
Notes:
- Uses
unified_planning.shortcuts.OptimizationValidator. - Ensures that optimized goals are reachable and valid.
Example Workflow
# Load and transform domain
optimized_domain = load_and_transform_domain("domain.pddl")
# Optimize problem
optimized_problem = optimize_problem("task01.pddl", optimized_domain)
# Validate optimization
if not validate_optimization(optimized_problem):
raise ValueError("Optimized problem is invalid")
# Save optimized problem
save_optimized_problem(optimized_problem, "optimized_task01.pddl")
Notes
- This skill set enhances the efficiency of PDDL planning problems through optimization.
- Designed for research in automated planning improvement.