COINCIDENCE ALGORITHM · LIVE RESEARCH ENVIRONMENT

COIN Live
Research Labs

A dedicated experimental environment for studying coincidence-based optimization—not a teaching-tool directory. Run reproducible benchmarks, inspect learning dynamics and compare permutation algorithms under explicit evaluation budgets.

Single + Multi-objectiveEqual evaluation budgetsCOIN · EHBSA · NHBSA · GANumPy / Numba

Choose a research problem

Each lab exposes the representation, evaluator, algorithm parameters, learning progress and result evidence appropriate to its problem.

FOUNDATIONAL LAB · EDGE LEARNING

Knight’s Tour

Observe directed edge probabilities, open and closed tours, Pareto depth and unique-solution storage on the original COIN demonstration problem.

Open live experiment →
PARALLEL COIN · DEPENDENCY-GUIDED DECOMPOSITION

Parallel COIN

Inspect candidate, position and tiled decomposition; compare Python, NumPy, Numba and OpenCL measurements with explicit pilot boundaries.

Open research slides →GitHub →
COIN · OBJECTIVE × REPRESENTATION

Objective Representation Study

Read the complete horizontal research deck: objectives, COIN variants, matched scalar rankings, convergence and Pareto evidence.

Open complete deck → ROSE · RELATIONAL ORDER

ROSE Evidence Deck

Compare relational-order variants across Taillard difficulty strata with RPD, paired ranks, convergence and multi-objective frontiers.

Open ROSE slides → SUDOKU · ROSE RELATIONAL-ORDER STUDY

ROSE on Sudoku

Follow the live one-seed checkpoint puzzle by puzzle: exact Sudoku grids, scaled populations, evaluations-to-solution and 800-generation convergence.

Open live Sudoku slides → CARE-EDA · ADAPTIVE REPRESENTATION

CARE-EDA

See how edge, position and relative-order generators receive adaptive offspring allocation from population feedback.

Open CARE slides → NB-COIN-E · OBJECTIVE SPECIALISTS

NB-COIN-E

Inspect fixed objective-specialist mixtures, scalar-to-Pareto allocation ratios and the current EX-COIN evidence.

Open NB-COIN-E slides → 象棋 · 90-POSITION PERMUTATION

Xiàngqí Knight’s Tour

Revisit the recorded 2011 open and closed tours on a 9×10 Chinese-chess board; compare SNE-COIN learning with randomized Warnsdorff under explicit budgets.

Open Xiàngqí lab → START-NODE EDGE COIN · DIVIDE & CONQUER

64×64 Closed Knight’s Tour

Learn endpoint-constrained 8×8 tours one subboard at a time, fold the macro route back, and verify all 4,096 moves including closure.

Open large-board lab → SCHEDULING · TAILLARD · PARETO

Permutation Flow Shop

Compare COIN variants, histogram models and pymoo algorithms with shared evaluators and identical objective-evaluation budgets.

Open benchmark → BIN PACKING · BPPLIB · DECODER

COIN/ROSE Load Partitioning

Decode a pure item permutation by opening a new sack only when the next item exceeds capacity, then compare bin count, unused capacity and load balance.

Open packing benchmark → PARALLEL MACHINES · FIXED PARTITIONS · TARDINESS

Parallel-machine Tardiness

Encode a fixed number of machine partitions as separator tokens and study total tardiness, maximum tardiness, makespan and load balance together.

Open tardiness benchmark → MAXIMIZATION · LOLIB · PERMUTATION

Linear Ordering Problem

Maximize weighted forward preferences on standard LOLIB Input/Output and Stanford GraphBase instances under equal evaluation budgets.

Open LOP benchmark → RCPSP · PRECEDENCE DAG · PARETO

Precedence-Constrained Scheduling

Decode learned priority lists with a Serial Schedule Generation Scheme and compare project duration, tardiness, throughput and resource leveling.

Open project scheduling lab → DNA/RNA · CONFLICT GRAPH · SUBSET SELECTION

DNA/RNA Helix Selection

Learn permutation priorities over competing stems and decode them into compatible helix subsets for multi-objective analysis.

Open helix-selection lab → ROUTING · TIME WINDOWS

TSPTW

Explore travel cost, makespan, waiting and lateness as single or competing objectives.

Run routing study → PERMUTATION · PRIME PRODUCT

Sudoku Optimization

Compare legacy, prime-product, hierarchical and multi-objective fitness signals with controllable obvious-fill preprocessing.

Open Sudoku lab → SUBSET SELECTION · BIOINFORMATICS

RNA Secondary Structure

Select compatible helices, optimize proxy and physical objectives, and inspect NAVIEW, RNApuzzler, circular and 3D topology views.

Open RNA lab → SOLUTION ARCHIVE · THREE.JS

Infinite Knight’s Tour

A continuous visual study generated from verified tours stored by the COIN solution service.

Enter visualization →
Research infrastructure, separated from teaching utilities

This directory now contains only laboratories that run COIN, ROSE, their EDA relatives, or verified COIN solution archives. Network and general teaching simulations have returned to Tools & Innovation.

Queue control