This project called ALO is designed for PCB Assembly line optimization which contains the algorithm implementation of the author and other works. The algorithm provided by ALO is applicable to the optimization of the assembly process of linear-aligned surface mount machine (SMM), such as SM482, and the production lines they form. The built-in optimization results provided by Hanwha Corporation can also be analyzed through ALO. The main algorithms were all designed and completed by the Assembly Process Optimization Research Group of the Institute of Intelligent Control and System, Harbin Institute of Technology. The implementation of the algorithms can be found in the papers in the [5 References] section.

1 File Structure

base_optimizer : optimizer of single surface mounter
- optimizer_common.py : basic common file for optimizer
- optimizer_interface.py : interface for all base optimizer
- result_analysis.py : result analysis of assembly process drawing and time calculation
- smt_aggregation.py : aggregation algorithm
- smt_celldivision.py : cell division algorithm
- smt_feederpriority.py : feeder priority algorithm
- smt_hybridgenetic.py : hybrid genetic algorithm
- smt_mathmodel.py : math model solver
- smt_scanbased.py : scan based algorithm
- smt_twophase.py : two phase algorithm
- smt_route.py : route schedule algorithm
data : pcb file directory
opt : training and test data for estimator directory
model : model parameter file
result : optimization result file
dataloader.py : load pcb file from *.txt
estimator.py : assembly time estimator including heuristic-based/neural-based/re-configuration-based/svr-based/metric-based estimator
generator.py : generate data for training the time estimator
lineopt_genetic.py : genetic algorithm
lineopt_model.py : math model
lineopt_hypergenetic.py : hyper-heuristic algorithm
lineopt_reconfiguration.py : reconfiguration algorithm
optimizer.py : main file
ReadMe.md : this file

2 Package Requirement

tqdm
gurobipy (license requirement)
sklearn（scikit-learn）
torch（PyTorch）
joblib
numpy
pandas
matplotlib
openpyxl

3 File Form

3.1 Part Section

part : part name
fdr : feeder type
nz : nozzle type
fdn : upper limit of feeder
points : number of feeders in step list

3.2 Step Section

machine : assigned index of machine
x : x position
y : y position
r : angle
part : part name
fdr : feeder slot & part name
nz : anc hole & nozzle tpye
hd : head index
cy : cycle index
others are not used

4 Parameter

mode ： run mode - 0: read optimized result, 1: run optimization, 2: batch optimize all PCB files in \data directory with machine optimizer, 3: batch optimize all PCB files in data/LINE#x with line optimizer and x + 2 is the number of machines in line
filename : PCB file (default in \data directory)
comp_register : whether auto-register components from step data when missing
machine_number : Number of machines in production line (set to 1 for single machine optimization)
machine_optimizer : single machine optimizer (options: cell-division, feeder-priority, hybrid-genetic, aggregation, genetic-scanning, mip-model, two-phase)
line_optimizer : line optimizer (options: hyper-heuristic, genetic, reconfiguration)
feeder_limit : available feeders for auto-registration
save : wheter save optimization results

5 Reference

5.1 SMM Optimizer

Aggregation Algorithm : ASHAYERI J, MA N, SOTIROV R. An aggregated optimization model for multihead SMD placements[J]. Computers & Industrial Engineering, 2011, 60(1): 99-105.
Cell Division Algorithm : LI Z, YU X, QIU J, et al. Cell division genetic algorithm for component allocation optimization in multi-functional placers[J]. IEEE Transactions on Industrial Informatics, 2022, 18(1): 559-570.
Math Model & Feeder Priority Algorithm : LU G, YU X, SUN H, et al. A Scan-Based Hierarchical Heuristic Optimization Algorithm for PCB Assembly Process[J]. IEEE Transactions on Industrial Informatics. 2024, 20(3): 3609-3618.
Hybrid Genetic Algorithm : GUO S, GENG F, TAKAHASHI K, et al. A MCVRP-based model for PCB assembly optimisation on the beam-type placement machine[J]. International Journal of Production Research, 2018, 57(18): 5874-5891.
Scan-Based Algorithm : SUN D S, LEE T E, KIM K H. Component allocation and feeder arrangement for a dual-gantry multi-head surface mounting placement tool[J]. International Journal of Production Economics, 2005, 95(2): 245-264.
Two Phase Algorithm : LU G, LI Z, SUN H, et al. A Two-Phase PCBA Optimization with ILP Model and Heuristic for a Beam Head Placement Machine[J]. IEEE Transactions on Industrial Informatics, 2024, 20(12):13612-13621.
Route Schedule Algorithm : From all above

5.2 Line Optimizer

Genetic Line Algorithm : GUO S, TAKAHASHI K, MORIKAWA K, et al. An integrated allocation method for the PCB assembly line balancing problem with nozzle changes[J]. International Journal of Advanced Manufacturing Technology, 2012, 62(1): 351-369.
Math Model & Hyper-Heuristic Line algorithm : LU G, GAO H, LI Z, et al. Hyper-Heuristic Optimization Using Multifeature Fusion Estimator for PCB Assembly Lines with Linear-Aligned-Heads Surface Mounters[J]. IEEE Transactions on Cybernetics. DOI:10.1109/TCYB.2025.3556512.
Re-Configuration Line algorithm : TÓTH A, KNUUTILA T, NEVALAINEN O S. Reconfiguring flexible machine modules of a PCB assembly line[J]. Production Engineering, 2010, 4(1): 85-94.

5.3 Estimator

Neural Estimator : VAINIO F, MAIER M, KNUUTILA T, et al. Estimating printed circuit board assembly times using neural networks[J]. International Journal of Production Research, 2010, 48(8): 2201-2218.
Re-Configuration Estimator : TÓTH A, KNUUTILA T, NEVALAINEN O S. Reconfiguring flexible machine modules of a PCB assembly line[J]. Production Engineering, 2010, 4(1): 85-94.
SVR Estimator : LI D, CHEN S, CHIONG R, et al. Predicting the printed circuit board cycle time of surface-mount-technology production lines using a symbiotic organism searchbased support vector regression ensemble[J]. International Journal of Production Research, 2020, 59(23): 7246-7265.
Heuristic Estimator & Metric Estimator : No reference

6 Copyright

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so. The original algorithm concept is the work of the cited authors.

ReadMe.md Unescape Escape

Assembly Line Optimizer