Return to Answer

import numpy as np
import sympy as sp
import random
import time
SIZE = 30

random.seed(0)

def gen_diag():
    return [random.randint(0, 1) for i in range(SIZE*2 - 1)]

def diag_to_mat(diag):
    return [diag[a:a+SIZE] for a in range(SIZE-1, -1, -1)]

def diag_to_det(diag):
    matrix = diag_to_mat(diag)
    return np.linalg.det(matrix)

def best_diags(diags, keep):
    return sorted(diags, key=diag_to_det, reverse=True)[:keep]

def improve(diag):
    old_diag = diag
    really_old_diag = []
    while really_old_diag != old_diag:
        really_old_diag = old_diag
        for flip_at in range(SIZE * 2 - 1):
            new_diag = old_diag[:]
            new_diag[flip_at] ^= 1
            old_diag = max(old_diag, new_diag, key=diag_to_det)
    return old_diag

overall_best_score = 0
time.clock()
while time.clock() < 500:
    best = improve(gen_diag())
    best_score = diag_to_det(best)
    if best_score > overall_best_score:
        overall_best_score = best_score
        overall_best = best
        print(time.clock(), sp.Matrix(diag_to_mat(overall_best)).det(), ''.join(map(str,overall_best)))


mat = diag_to_mat(overall_best)

sym_mat = sp.Matrix(mat)

print(overall_best)
print(sym_mat.det())

import numpy as np
import sympy as sp
import random
import time
SIZE = 30

random.seed(0)

def gen_diag():
    return [random.randint(0, 1) for i in range(SIZE*2 - 1)]

def diag_to_mat(diag):
    return [diag[a:a+SIZE] for a in range(SIZE-1, -1, -1)]

def diag_to_det(diag):
    matrix = diag_to_mat(diag)
    return np.linalg.det(matrix)

def improve(diag):
    old_diag = diag
    really_old_diag = []
    while really_old_diag != old_diag:
        really_old_diag = old_diag
        for flip_at in range(SIZE * 2 - 1):
            new_diag = old_diag[:]
            new_diag[flip_at] ^= 1
            old_diag = max(old_diag, new_diag, key=diag_to_det)
    return old_diag

overall_best_score = 0
time.clock()
while time.clock() < 500:
    best = improve(gen_diag())
    best_score = diag_to_det(best)
    if best_score > overall_best_score:
        overall_best_score = best_score
        overall_best = best
        print(time.clock(), sp.Matrix(diag_to_mat(overall_best)).det(), ''.join(map(str,overall_best)))


mat = diag_to_mat(overall_best)

sym_mat = sp.Matrix(mat)

print(overall_best)
print(sym_mat.det())

Matrices found with this determinant, value of diagonal from bottom left to upper right.

01000100101101000011100111011101000100101101000011100111011
01011101110011100001011010010001011101110011100001011010010
01100001000111011101001110100101100001000111011101001110100

Matrices found with this determinant, given as value of diagonal from bottom left to upper right:

01000100101101000011100111011101000100101101000011100111011
01011101110011100001011010010001011101110011100001011010010
01100001000111011101001110100101100001000111011101001110100
01110100111010010110000100011101110100111010010110000100011
01011101110001000011010010111001011101110001000011010010111
01000101100010110100111101110001000101100010110100111101110
01000100101101000011100111011101000100101101000011100111011

This is hill climbing, as straightforward as can be. Final determinant calculation performed using SymPy to enusure an exact result. Both matrices found are circulant.

Diagonals:

01000100101101000011100111011101000100101101000011100111011
01011101110011100001011010010001011101110011100001011010010

The former, as a matrix:

This is hill climbing, as straightforward as can be. Final determinant calculation performed using SymPy to enusure an exact result. All matrices found with this determinant are circulant.

Matrices found with this determinant, value of diagonal from bottom left to upper right.

01000100101101000011100111011101000100101101000011100111011
01011101110011100001011010010001011101110011100001011010010
01100001000111011101001110100101100001000111011101001110100

The first one, as a matrix: