Merge pull request #302 from HenriqueAssumpcao/feature/alphaevolve_math_problems

Add all AlphaEvolve mathematical problems

Author: codelion

examples/alphaevolve_math_problems/README.md

@@ -0,0 +1,23 @@
+## AlphaEvolve mathematical problems
+
+This folder contains the necessary evaluator and initial program files for all of the 14 problems from [AlphaEvolve's Appendices A and B](https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/AlphaEvolve.pdf). The circle packing problem was among the first implemented by OpenEvolve. This folder implements the remaining 13 problems. 
+
+Not all problems take the form of a maximization problem, however in order to make this problem set more standardized we chose to **make it so all evaluator files are aiming to maximize the target metric**. We achieve by some straightforward algebraic manipulation, but this can be easily edited and changed if the user finds it necessary.
+
+The problem from Appendix A is the following:
+- [Matrix multiplication](matmul): obtain a faster algorithm for multiplying two matrices of sizes $m \times n$ and $n \times p$ (c.f. Appendix A).
+
+The remaining problems are from Appendix B:
+1. [First autocorrelation inequality](first_autocorr_ineq): Construct a nonnegative step function $f:\mathbb{R} \mapsto \mathbb{R}$ to improve an upper bound on a constant related to the autoconvolution of $f$ (c.f. Appendix B.1.).
+2. [Second autocorrelation inequality](second_autocorr_ineq): Construct a nonnegative step function $f:\mathbb{R} \mapsto \mathbb{R}$ to improve a lower bound on a constant related to the norm of the autoconvolution of $f$ (c.f. Appendix B.2.).
+3. [Third autocorrelation inequality](third_autocorr_ineq): Construct a nonnegative step function $f:\mathbb{R} \mapsto \mathbb{R}$ to improve an upper bound on a constant related to th absolute value of the autoconvolution of $f$ (c.f. Appendix B.3.).
+4. [An uncertainty inequality](uncertainty_ineq): Construct a function $f:\mathbb{R} \mapsto \mathbb{R}$ to obtain an upper bound on a constant related to $f$ and its fourier transform. (c.f. Appendix B.4.).
+5. [Erdos minimum overlap problem](erdos_min_overlap): Construct a nonnegative step function $f:\mathbb{R} \mapsto \mathbb{R}$ satisfyind some special properties to improve an upper bound on a constant that controls the asymptotics of the Minimum Overlap Problem (c.f. Appendix B.5.).
+6. [Sums and differences of finite sets](sums_diffs_finite_sets): Construct a set of nonnegative integers $U$ satisfying some special properties to improve a lower bound to a constant related to sums and differences of finite sets (c.f. Appendix B.6.).
+7. [Packing unit regular hexagons inside a regular hexagon](hexagon_packing): Place $n$ disjoint unit regular hexagons inside a larger regular hexagon, minimizing the side length of the outer hexagon. We consider the case where $n = 11$ and $n = 12$ (c.f. Appendix B.7.).
+8. [Minimizing the ratio of maximum to minimum distance](minimizing_max_min_dist): Place $n$ $d$-dimensional points in order to minimize the ratio between the maximum and minimum pairwise distances. We consider the cases where $n=16,d=2$ and $n=14,d=3$ (c.f. Appendix B.8.).
+9. [The Heilbronn problem for triangles](heilbronn_triangle): Place $n$ points on or inside a triangle with unit area so that the area of the smallest triangle formed by these points is maximized. We consider the case where $n = 11$ (c.f. Appendix B.9.).
+10. [The Heilbronn problem for convex regions](heilbronn_convex): Place $n$ points on or inside a convex region with unit area so that the area of the smallest triangle formed by these points is maximized. We consider the case where $n = 13$ and $n=14$ (c.f. Appendix B.10.).
+11. [Kissing number in dimension 11](kissing_number): Increase the lower bound on the $11$-dimensional kissing number, i.e., the number of disjoint unit spheres that can be packed tangent to a given unit sphere (c.f. Appendix B.11.).
+12. [Packing circles inside a unit square to maximize sum of radii](../circle_packing): Place $n$ disjoint circles inside a unit square so as to maximize the sum of their radii (c.f. Appendix B.12.).
+13. [Packing circles inside a rectangle of perimeter 4 to maximize sum of radii](circle_packing_rect):  Place $n$ disjoint circles inside a rectangle of perimeter $4$ so as to maximize the sum of their radii (c.f. Appendix B.13.).

examples/alphaevolve_math_problems/circle_packing_rect/config.yaml

@@ -0,0 +1,62 @@
+# Evolution settings
+max_iterations: 100
+checkpoint_interval: 10
+parallel_evaluations: 1
+
+# LLM configuration
+llm:
+  api_base: "https://api.openai.com/v1"  # Or your LLM provider
+  models:
+    - name: "gpt-4"
+      weight: 1.0
+  temperature: 0.7
+  max_tokens: 4000
+  timeout: 120
+
+# Database configuration (MAP-Elites algorithm)
+database:
+  population_size: 40
+  num_islands: 5
+  migration_interval: 40
+  feature_dimensions:  # MUST be a list, not an integer
+    - "score"
+    - "complexity"
+
+# Evaluation settings
+evaluator:
+  timeout: 360
+  max_retries: 3
+
+# Prompt configuration
+prompt:
+  system_message: |
+    SETTING:
+    You are an expert computational geometer and optimization specialist with deep expertise in circle packing problems, geometric optimization algorithms, and constraint satisfaction.
+    Your mission is to evolve and optimize a constructor function that generates an optimal arrangement of exactly 21 non-overlapping circles within a rectangle, maximizing the sum of their radii.
+
+    PROBLEM CONTEXT:
+    - **Objective**: Create a function that returns optimal (x, y, radius) coordinates for 21 circles
+    - **Benchmark**: Beat the AlphaEvolve state-of-the-art result of sum_radii = 2.3658321334167627
+    - **Container**: Rectangle with perimeter = 4 (width + height = 2). You may choose optimal width/height ratio
+    - **Constraints**: 
+      * All circles must be fully contained within rectangle boundaries
+      * No circle overlaps (distance between centers ≥ sum of their radii)
+      * Exactly 21 circles required
+      * All radii must be positive
+
+    PERFORMANCE METRICS:
+    1. **sum_radii**: Total sum of all 21 circle radii (PRIMARY OBJECTIVE - maximize)
+    2. **combined_score**: sum_radii / 2.3658321334167627 (progress toward beating benchmark)  
+    3. **eval_time**: Execution time in seconds (keep reasonable, prefer accuracy over speed)
+
+    TECHNICAL REQUIREMENTS:
+    - **Determinism**: Use fixed random seeds if employing stochastic methods for reproducibility
+    - **Error handling**: Graceful handling of optimization failures or infeasible configurations
+    - **Memory efficiency**: Avoid excessive memory allocation for distance matrix computations
+    - **Scalability**: Design with potential extension to different circle counts in mind
+      
+  num_top_programs: 3
+  num_diverse_programs: 2
+
+# Logging
+log_level: "INFO"

examples/alphaevolve_math_problems/circle_packing_rect/evaluator.py

@@ -0,0 +1,111 @@
+# ===--------------------------------------------------------------------------------------===#
+#
+# This file implements the evaluator for the circle packing problem on a rectangle
+# of perimeter 4.
+#
+# ===--------------------------------------------------------------------------------------===#
+#
+# Some of the code in this file is adapted from:
+#
+# google-deepmind/alphaevolve_results:
+# Licensed under the Apache License v2.0.
+#
+# ===--------------------------------------------------------------------------------------===#
+
+import time
+import numpy as np
+import sys
+import os
+from importlib import __import__
+
+BENCHMARK = 2.3658321334167627
+NUM_CIRCLES = 21
+TOL = 1e-6
+
+
+def minimum_circumscribing_rectangle(circles: np.ndarray):
+    """Returns the width and height of the minimum circumscribing rectangle.
+
+    Args:
+    circles: A numpy array of shape (num_circles, 3), where each row is of the
+        form (x, y, radius), specifying a circle.
+
+    Returns:
+    A tuple (width, height) of the minimum circumscribing rectangle.
+    """
+    min_x = np.min(circles[:, 0] - circles[:, 2])
+    max_x = np.max(circles[:, 0] + circles[:, 2])
+    min_y = np.min(circles[:, 1] - circles[:, 2])
+    max_y = np.max(circles[:, 1] + circles[:, 2])
+    return max_x - min_x, max_y - min_y
+
+
+def validate_packing_radii(radii: np.ndarray) -> None:
+    n = len(radii)
+    for i in range(n):
+        if radii[i] < 0:
+            raise ValueError(f"Circle {i} has negative radius {radii[i]}")
+        elif np.isnan(radii[i]):
+            raise ValueError(f"Circle {i} has nan radius")
+
+
+def validate_packing_overlap_wtol(circles: np.ndarray, tol: float = 1e-6) -> None:
+    n = len(circles)
+    for i in range(n):
+        for j in range(i + 1, n):
+            dist = np.sqrt(np.sum((circles[i, :2] - circles[j, :2]) ** 2))
+            if dist < circles[i, 2] + circles[j, 2] - tol:
+                raise ValueError(
+                    f"Circles {i} and {j} overlap: dist={dist}, r1+r2={circles[i,2]+circles[j,2]}"
+                )
+
+
+def validate_packing_inside_rect_wtol(circles: np.array, tol: float = 1e-6) -> None:
+    width, height = minimum_circumscribing_rectangle(circles)
+    if width + height > (2 + tol):
+        raise ValueError("Circles are not contained inside a rectangle of perimeter 4.")
+
+
+def evaluate(program_path: str):
+    try:
+        abs_program_path = os.path.abspath(program_path)
+        program_dir = os.path.dirname(abs_program_path)
+        module_name = os.path.splitext(os.path.basename(program_path))[0]
+
+        circles = None
+        eval_time = 0
+        try:
+            sys.path.insert(0, program_dir)
+            program = __import__(module_name)
+
+            start_time = time.time()
+            circles = program.circle_packing21()
+            end_time = time.time()
+            eval_time = end_time - start_time
+        except Exception as err:
+            raise err
+        finally:
+            if program_dir in sys.path:
+                sys.path.remove(program_dir)
+
+        if not isinstance(circles, np.ndarray):
+            circles = np.array(circles)
+
+        if circles.shape != (NUM_CIRCLES, 3):
+            raise ValueError(
+                f"Invalid shapes: circles = {circles.shape}, expected {(NUM_CIRCLES,3)}"
+            )
+
+        validate_packing_radii(circles[:, -1])
+        validate_packing_overlap_wtol(circles, TOL)
+        validate_packing_inside_rect_wtol(circles, TOL)
+
+        radii_sum = np.sum(circles[:, -1])
+
+        return {
+            "radii_sum": float(radii_sum),
+            "combined_score": float(radii_sum / BENCHMARK),
+            "eval_time": float(eval_time),
+        }
+    except Exception as e:
+        return {"combined_score": 0.0, "error": str(e)}

examples/alphaevolve_math_problems/circle_packing_rect/initial_program.py

@@ -0,0 +1,22 @@
+# EVOLVE-BLOCK-START
+import numpy as np
+
+
+def circle_packing21() -> np.ndarray:
+    """
+    Places 21 non-overlapping circles inside a rectangle of perimeter 4 in order to maximize the sum of their radii.
+
+    Returns:
+        circles: np.array of shape (21,3), where the i-th row (x,y,r) stores the (x,y) coordinates of the i-th circle of radius r.
+    """
+    n = 21
+    circles = np.zeros((n, 3))
+
+    return circles
+
+
+# EVOLVE-BLOCK-END
+
+if __name__ == "__main__":
+    circles = circle_packing21()
+    print(f"Radii sum: {np.sum(circles[:,-1])}")

examples/alphaevolve_math_problems/circle_packing_rect/requirements.txt

@@ -0,0 +1 @@
+numpy

examples/alphaevolve_math_problems/erdos_min_overlap/config.yaml

@@ -0,0 +1,60 @@
+# Evolution settings
+max_iterations: 200
+checkpoint_interval: 10
+parallel_evaluations: 1
+
+# LLM configuration
+llm:
+  api_base: "https://api.openai.com/v1"  # Or your LLM provider
+  models:
+    - name: "gpt-4"
+      weight: 1.0
+  temperature: 0.7
+  max_tokens: 4000
+  timeout: 120
+
+# Database configuration (MAP-Elites algorithm)
+database:
+  population_size: 40
+  num_islands: 5
+  migration_interval: 40
+  feature_dimensions:  # MUST be a list, not an integer
+    - "score"
+    - "complexity"
+
+# Evaluation settings
+evaluator:
+  timeout: 360
+  max_retries: 3
+
+# Prompt configuration
+prompt:
+  system_message: |
+    SETTING:
+    You are an expert in harmonic analysis, numerical optimization, and AI-driven mathematical discovery.
+    Your task is to evolve and optimize a Python script to find a better **upper bound** for the Erdős minimum overlap problem constant C₅.
+    
+    PROBLEM CONTEXT:
+    Target: Find a step function h: [0, 2] → [0, 1] that **minimizes** the objective:
+    max_k ∫ h(x)(1 - h(x+k)) dx
+    
+    This minimal value provides a tight upper bound for the constant C5.
+    
+    Current best known upper bound: C5 ≤ 0.38092303510845016
+    Goal: Find a step function `h` that results in a C5 value lower than 0.38092303510845016.
+    
+    CONSTRAINTS:
+    1. The function `h` must have values in the range [0, 1].
+    2. The integral of h(x) over [0, 2] must be exactly 1.
+    
+    PERFORMANCE METRICS:
+    - c5_bound: The bound found by the program.
+    - combined_score: 0.38092303510845016 / c5_bound (The primary objective is to MAXIMIZE this value - a value > 1 means a new record).
+    - n_points: number of points used in the discretization.
+    - eval_time: evaluation time of the program.
+    
+  num_top_programs: 3
+  num_diverse_programs: 2
+
+# Logging
+log_level: "INFO"

examples/alphaevolve_math_problems/erdos_min_overlap/evaluator.py

@@ -0,0 +1,76 @@
+# ===--------------------------------------------------------------------------------------===#
+#
+# This file implements the evaluator for the erdos minimum overlap problem.
+#
+# ===--------------------------------------------------------------------------------------===#
+#
+# Some of the code in this file is adapted from:
+#
+# google-deepmind/alphaevolve_results:
+# Licensed under the Apache License v2.0.
+#
+# ===--------------------------------------------------------------------------------------===#
+
+import sys
+import os
+from importlib import __import__
+import time
+import numpy as np
+
+# Known bounds
+BENCHMARK = 0.38092303510845016
+
+
+def verify_c5_solution(h_values: np.ndarray, c5_achieved: float, n_points: int):
+    """Verifies the C5 upper bound solution."""
+
+    if h_values.shape != (n_points,):
+        raise ValueError(f"Expected h shape ({n_points},), got {h_values.shape}")
+
+    # Verify h(x) in [0, 1] constraint
+    if np.any(h_values < 0) or np.any(h_values > 1):
+        raise ValueError(f"h(x) is not in [0, 1]. Range: [{h_values.min()}, {h_values.max()}]")
+
+    # Verify integral of h = 1 constraint
+    dx = 2.0 / n_points
+    integral_h = np.sum(h_values) * dx
+    if not np.isclose(integral_h, 1.0, atol=1e-3):
+        raise ValueError(f"Integral of h is not close to 1. Got: {integral_h:.6f}")
+
+    # Re-calculate the C5 bound using np.correlate
+    j_values = 1.0 - h_values
+    correlation = np.correlate(h_values, j_values, mode="full") * dx
+    computed_c5 = np.max(correlation)
+
+    # Check for consistency
+    if not np.isclose(computed_c5, c5_achieved, atol=1e-4):
+        raise ValueError(f"C5 mismatch: reported {c5_achieved:.6f}, computed {computed_c5:.6f}")
+
+
+def evaluate(program_path: str):
+    try:
+        abs_program_path = os.path.abspath(program_path)
+        program_dir = os.path.dirname(abs_program_path)
+        module_name = os.path.splitext(os.path.basename(program_path))[0]
+
+        try:
+            sys.path.insert(0, program_dir)
+            program = __import__(module_name)
+            start_time = time.time()
+            h_values, c5_bound, n_points = program.run()
+            end_time = time.time()
+            eval_time = end_time - start_time
+        finally:
+            if program_dir in sys.path:
+                sys.path.remove(program_dir)
+
+        verify_c5_solution(h_values, c5_bound, n_points)
+
+        return {
+            "c5_bound": float(c5_bound),
+            "combined_score": BENCHMARK / float(c5_bound),
+            "n_points": int(n_points),
+            "eval_time": float(eval_time),
+        }
+    except Exception as e:
+        return {"combined_score": 0.0, "error": str(e)}