Skip to content

OR-Tools Examples

solvOR implementations of Google OR-Tools examples.

These examples demonstrate how to solve the same problems using solvOR's pure Python solvers. The goal is minimal code changes from the original OR-Tools examples while using solvOR's API.

All benchmarks run on the same machine with Python 3.13. Times include Python startup overhead.

Linear Solver

LP and MIP problems using solve_lp, solve_milp, solve_hungarian, solve_bin_pack.

Problem solvOR Time OR-Tools Time Ratio
Basic LP (2 vars, 1 constraint) simple_lp_program.py 70 ms original 62 ms 1.1x
Basic MIP (2 int vars) simple_mip_program.py 68 ms original 79 ms 0.9x
LP with 3 vars, 3 constraints linear_programming_example.py 70 ms original 64 ms 1.1x
MIP with 3 int vars integer_programming_example.py 75 ms original 73 ms 1.0x
GLOP solver demo basic_example.py 68 ms original 65 ms 1.1x
Stigler diet (77 foods, 9 nutrients) stigler_diet.py 79 ms original 63 ms 1.3x
Worker-task assignment (5x4) assignment_mip.py 71 ms original 70 ms 1.0x
Bin packing (11 items) bin_packing_mip.py 70 ms original 71 ms 1.0x
Multiple knapsack (15 items, 5 bins)* multiple_knapsack_mip.py 2.8 s original 109 ms 26x
Group-constrained assignment (12x6) assignment_groups_mip.py 81 ms original 84 ms 1.0x
GAP with task sizes (10x8) assignment_task_sizes_mip.py 381 ms original 88 ms 4.3x
Team-constrained assignment (6x4) assignment_teams_mip.py 80 ms original 79 ms 1.0x
MIP with variable array (5 vars) mip_var_array.py 131 ms original 91 ms 1.4x

CP-SAT

Constraint programming problems using Model and solve_job_shop.

Problem solvOR Time OR-Tools Time Ratio
Simple CSP (3 vars, x!=y) simple_sat_program.py 82 ms original 436 ms 0.2x
N-Queens (8x8, 92 solutions) nqueens_sat.py 289 ms original 443 ms 0.7x
Nurse scheduling (4x3x3) nurses_sat.py 122 ms original 416 ms 0.3x
Job shop (3 jobs, 3 machines) minimal_jobshop_sat.py 89 ms original 405 ms 0.2x
All solutions (3 vars, 18 sols) search_for_all_solutions_sample_sat.py 95 ms original 434 ms 0.2x
Stop after N (5 solutions) stop_after_n_solutions_sample_sat.py 109 ms original 447 ms 0.2x

Graph

Graph algorithms using max_flow, solve_hungarian.

Problem solvOR Time OR-Tools Time Ratio
Max flow (5 nodes, 9 arcs) simple_max_flow_program.py 96 ms original 162 ms 0.6x
Linear sum assignment (4x4) assignment_linear_sum_assignment.py 127 ms original 196 ms 0.7x

Algorithms

Specialized algorithms using solve_knapsack.

Problem solvOR Time OR-Tools Time Ratio
Simple knapsack (18 items) simple_knapsack_program.py 120 ms original 83 ms 1.4x
Knapsack (50 items)* knapsack.py 91 ms original 106 ms 0.9x

* knapsack.py: Same objective value (7534) but solvOR solution may exceed capacity. Works correctly when values=weights.

Notes

  • solvOR is pure Python, designed for learning and prototyping
  • OR-Tools has a C++ backend, optimized for production workloads
  • For small problems, Python startup time dominates (both ~60-70ms)
  • For larger problems, expect OR-Tools to be 10-100x faster
  • Results verified to match between implementations (except where noted)

Source Code