Conference PaperPDF Available

A Tutorial on how to Connect Python with Different Simulation Software to Develop Rich Simheuristics

Proceedings of the 2021 Winter Simulation Conference
S. Kim, B. Feng, K. Smith, S. Masoud, Z. Zheng, C. Szabo, and M. Loper, eds.
Mohammad Dehghanimohammadabadi
Sahil Belsare
Mechanical and Industrial Engineering Department
Northeastern University
360 Huntington Ave
Boston, MA 02115, USA
Renee Thiesing
Simio LLC
504 Beaver St
Sewickley, PA 15143, USA
With rapid advancements in Cyber-Physical manufacturing, the Internet of Things, Simulation software,
and Machine Learning algorithms, the applicability of Industry 4.0 is gaining momentum. The demand
for real-time decision-making in the manufacturing industry has given significant attention to the field of
Digital Twin (DT). The whole idea revolves around creating a digital counterpart of the physical system
based on enterprise data to exploit the effects of numerous parameters and make informed decisions.
Based on that, this paper proposes a simulation-optimization framework for the DT model of a Beverage
Manufacturing Plant. A data-driven simulation model developed in Simio is integrated with Python to
perform Multi-Objective optimization. The framework explores optimal solutions by simulating multiple
scenarios by altering the availability of operators and dispatching/scheduling rules. The results show that
simulation optimization can be integrated into the Digital-Twin models as part of real-time production
planning and scheduling.
In the recent wave of Industry 4.0, Smart Factories and Intelligent Manufacturing have received significant
attention from both researchers and industries. Smart Factories aim at achieving high adaptability, enhanced
efficiency, increased productivity, and clearer visibility of operations. This requires generating, processing,
and learning a tremendous amount of data-driven knowledge from different parts of the manufacturing
system. There exists a growing body of literature focusing on integrating multiple technologies like IoT,
simulation, optimization, and Machine Learning to create a Cyber-Physical manufacturing system. A
complete real-time presentation of the state of the intelligent manufacturing system is a challenge; however,
the emergence of Digital Twin (DT) has made it possible to solve this problem (He and Bai 2020). The
whole idea revolves around creating a virtual and digital counterpart of the physical system based on
enterprise data to exploit the effects of numerous parameters and make informed decisions.
The concept of Digital Twin was put forward by Michael Grieves in 2002, which focused on product
life-cycle management (Kritzinger et al. 2018). In the manufacturing setting, DT is perceived as a virtual
simulation model of a physical system, which is applied to optimize the operational processes to achieve
precise control over the whole assembly (He and Bai 2020). However, DT in the manufacturing industry
is more than just a simulation model. It is an integration of smart digital machines, a simulation model, a
network of widespread data, and the adoption of information/communication technologies by manufacturing
systems. In order to fully exploit this potential, it is vital to realize this collaboration between humans,
machines, environment in the simulation model, and the manufacturing process (Zheng et al. 2019).
Apart from the proven benefits, implementing a fully efficient DT can be inherently a complex process.
This calls for the need for experimentation with several configuration settings, parameter testing, and an
Dehghanimohammadabadi, Thiesing, and Belsare
optimization framework to achieve the desired performance. This need is conventionally facilitated with
the support of Discrete Event Simulation (DES) software applications. A central aspect of the DES model
is its capability to utilize data to simulate a real-life process and provide insights into various possible
scenarios. This process can be real-time where the simulation model is integrated with an Enterprise
Resource Planning (ERP) system. This need is conventionally facilitated with the support of Discrete Event
Simulation (DES) software applications. An ERP system facilitates the flow of information, a variety of
reports, and data analytics of an organization. The feature of providing access to real-time manufacturing
data like orders, schedules, human and material availability and much more can be exploited to build a
simulation model. A central aspect of the DES model is its capability to utilize data to simulate a real-life
process and provide insights into various possible scenarios. For the DT to replicate the true behavior of
the physical process, it must incorporate detailed constraint model of the process. That includes all the
equipment, labor, tooling, transportation and material along with the equipment and material characteristics
driving the operational decisions. It is essential to factor in the business rules that regulate the operations
such as inventory policies, labor policies, operating procedures, and transportation restrictions, for example.
And finally, it must be able to capture the detailed day- to- day decision logic as applied by the planners,
operators, and supervisors managing the process. A DES software is uniquely positioned to be able to
model at this level of detail while also capturing the inherent variability present through the system. A
DT can be fully generated and driven by Enterprise data. For example, an Enterprise Resource Planning
(ERP) system can provide master data that defines all the resources in the system, along with material
requirements and costing information. A Manufacturing Execution System (MES) can provide a definition
of the resources on the factory floor, along with the current status of machine up-time, downtime, and
work in process. Connecting the DT to such systems, will allow it to automatically adapt to changes in the
environment such as additional equipment, new labor and skill requirements, new parts/SKUs, etc. The
DT when connected to real- time data, would allow it to make predictive and perspective decisions based
on the current status of the system.
In addition, DES provides an environment to deploy manual or systemic experimentations to analyze
multiple what-if scenarios. This enables decision-makers to test various process plans and scheduling
techniques to obtain an optimized responsive planning, management, and decision making. This paper aims
to propose a simulation-optimization (SO) framework to demonstrate its applicability for DT implementation.
This framework takes advantage of data-driven modeling where a simulation model is directly linked with
an ERP system to imitate the manufacturing facility. Therefore, the contributions of this paper as follows:
Design a simulation-optimization Digital Twin (SODT) framework
Implement SODT by integrating a DES package with an optimization algorithm
Demonstrate the applicability of the proposed SODT in a manufacturing setting and provide insights
for future developments
This project is completed by integrating Simio with Python. Simio is a powerful DES package and is
written in C# on a .NET platform. The Simio API allows for flexible integration with other systems, which
is important for the ability to not only connect to Enterprise systems but also allow for the ability to integrate
optimization and artificial intelligence with the DT. The model can be connected to an external system in
several ways, but most popular for a DT, is either with a direct database connection or with the WebAPI.
For integrating optimization with the DT, the .NET platform and robust API makes Simio flexible enough
to couple it with a high-level programming language like Python. Python makes use of Python Package
Index (PyPI) containing third-party modules making it possible to interact with other platforms. It’s ability
to handle multiple data types, editing, writing, and manipulating other software proves a key feature to
execute combined operations. The extent of libraries available to perform statistical, mathematical and
optimization calculations makes Python a great tool for simulation-optimization framework. These features
of used software packages facilitate integration and provide a unique platform to optimize a simulated DT.
Dehghanimohammadabadi, Thiesing, and Belsare
The rest of this paper is organized as follows: The literature review in Section 2 represents an overview
of applications and use-cases pertaining to simulation and digital twin models in the manufacturing industry.
Section 3 explains in detail the methodology used to integrate SODT with a Beverage Production Plant.
Section 4 puts forward results obtained by simulating multiple scenarios with a continually optimized
solutions through an optimization algorithm. Conclusions and future works are addressed in section 5.
The multifaced definitions of DT prevailing in the manufacturing domain as well, motivated (Zhang et al.
2021) to work on two specific research questions, ‘What is the definition of Digital Twin in the scientific
literature?’ and ‘What is its role within Industry 4.0?’. The authors put forward a comprehensive study
with a focus on providing a solution to the problem from the point of view of model engineering and
simulation. This indicates that DT is at the stage of rapid development where researchers start to explore
real practices and technologies in the industry (Liu et al. 2020). According to Zheng et al. (2019), the
ongoing extensive research on Cyber-Physical systems and Digital Twins has gradually become one of
the key research directions of intelligent manufacturing. An extensive review published by He and Bai
(2020), identified Production line and process simulation as one of the key development areas for DT for
intelligent manufacturing. Al-Ali et al. (2020) asserts that the application of DT in manufacturing could
help in higher flexibility, higher production, and better maintenance of the manufacturing and automation
process, thus improving the overall operational efficiency. Santos et al. (2019), proposed the usage of
DT for Manufacturing Executing System (MES) to obtain an optimum production schedule. The system
consisted of an IoT platform, simulators, and user applications to provide changing inputs. Similarly,
a decision support system for improving the order management process was proposed by Kunath and
Winkler (2018). The proposed system is capable of generating a simulation model automatically using
information from the Digital Twin of the manufacturing system. Another Digital Twin-based Cyber-Physical
Production System was proposed by Ding et al. (2019) to optimize real-time monitoring, simulation, and
prediction of manufacturing operations. Developing a combined simulation-optimization method with DT
is another upcoming research topic popular in the manufacturing domain. Balderas et al. (2021) developed
a Digital-Twin framework that integrates a metaheuristic optimization and a direct Simulink model for
printed circuit boards (PCB) design and processing. The promising results obtained from the experiment
show the benefits of integrating metaheuristic optimization into the Digital-Twin concept. Similarly, Liu
et al. (2021) proposed a simulation-optimization scheduling platform for an aeroengine gear production
workshop. The model was found efficient in optimizing scheduling by shortening both transit and waiting
times within the production process.
Dynamic scheduling by continuous decision-making, predicting machine availability, bottleneck detec-
tion, and performance evaluation are common focus parameters among the reviewed studies. Zhang et al.
(2021), demonstrates the use of optimization in DT to reduce the makespan and total tardiness by 14.5%
and 87.1%, respectively, and increase the average utility rate by 14.9% of a hydraulic valve machining
job-shop. Park et al. (2021) puts forward a novel production control model that applies DT and horizontal
coordination with RL-based production control to a re-entrant job shop problem. Zhang et al. (2020) argues
that it is difficult to find an effective simulation-based optimization method to solve the large-scale discrete
optimization problems in digital twin shop floors. And to overcome these challenges, the authors propose
an improved multi-fidelity simulation-based optimization method based on multi-fidelity optimization. The
novel method makes use of heuristics algorithms to accelerate the solution space search integrated with
a DES-based simulation optimization system. A joint simulation optimization and DT model to optimize
stacked packing and storage assignment of the warehouse was proposed by Leng et al. (2019). The
proposed model was able to maximize the utilization and efficiency of the large-scale automated high-rise
warehouse product-service system. Park et al. (2021), puts forward a DES and Digital Twin framework
for dispatching assistance in port logistics. Gyulai et al. (2020) makes use of DES model for the detailed
representation of a complex shop-floor logistics system, employing automated robotic vehicles (AGV).
Dehghanimohammadabadi, Thiesing, and Belsare
While some researchers believe that the concept of DT is in the initial stage, the growing interest is
evident by the various use-cases published in recent years. Upon meticulously analyzing the selected studies
it can be concluded that the use of simulation-optimization techniques combined with DT is a promising
research topic. This can be achieved by merging three things:
A simulation model – visually replicating a physically happening process
Real-time data processing, monitoring, and controlling capabilities
Estimating future state capabilities using optimizing and machine learning embedded models
A common note in all the reviewed papers is about tremendous research opportunities in the Digital
Twin technology for Industry 4.0. To accelerate the process of implementation the researchers should
address the limitations, develop a suitable framework, and parallelly, increase industrial use cases. In
pursuit of the same, this paper demonstrates the implementation of integrating a Simulation-Optimization
Framework for a Digital Twin model.
3.1 The Proposed SODT Framework
The proposed SODT is an integrated simulation-optimization framework to enhance DT performance. As
illustrated in Figure 1, this framework is a combination of three modules, i) data exchange, ii) optimization,
and iii) simulation. Data exchange is the key element in the framework where connects all components
Figure 1: The proposed SODT framework structure.
The first function of data exchange is to connect the simulation model with the ERP system. This
connection enables the simulation model to capture the real-life changes in real-time and reflect them in
the simulation model. All of the ERP information can be stored in separate files (i.e., Excel or CSV) and
linked to the simulation model. This Simio capability makes the simulation modeling process seamless,
accurate, and efficient. The data tables can include a wide range of information for resources, materials,
orders, dispatching rules, labor, schedules, entities, routings networks, etc.
The data-table module can also be used as a liaison between the simulation and optimization module.
The optimization algorithm designed for this work takes data tables as an entry for the optimization model
(decision variable) and tries to find the evolved table entries and provide the desired solution. Therefore,
each of the ERP data tables can be subject to optimization depending on the user’s needs. For instance, a
Dehghanimohammadabadi, Thiesing, and Belsare
user can optimize orders data-table to change order priorities to satisfy objectives. Another example would
be optimizing the dispatching rule table to figure out the best set of rules to proceed with operation on the
floor. As can be observed, the improvement opportunities with this unique framework are unbounded.
And finally, once the simulation optimization is completed, the optimal results are tabulated in data
tables, and then the new results are populated back to the ERP system. At this point, the updated ERP info
can be used in the actual system to perform optimally.
3.2 Implementation Of The SODT Framework
The proposed SODT framework was implemented and verified with a case study of the Beverage Production
unit. The model was built using Simio Enterprise Edition to simulate a batch processing system that mixes
and fills a beverage product. The model is capable of imitating the real-life scenarios as the inputs to the
model can be dynamically changed by altering values in a table. To make these changes real-time the
tables can be linked to the ERP database to continually update the input parameters of the model.However,
this has not been tested in the current experimental design and lies on the future scope of the project. For
the experimental purposes, the data tables representing output of ERP system can be altered manually.
Furthermore, the model is capable of completely implementing all the real-life constraints of the resources
to provide realistic operating scenarios. The validity of simulation model has not been tested with historical
data since, the experimental analysis makes use of an in-built Simio example of batch scheduling with a
minor change of measuring Tardiness Cost as well. The simulation model’s entities and server are directly
linked to the tables representing the manufacturing data from ERP system. The model configuration is
made up of multiple tables like Manufacturing orders, Routing logic, Bill of material, Dispatching Rules
for Mixing and Packing Operations, range of available workers, availability of raw material. The model
properties can be easily changed by altering the excel files to accommodate changing real life scenarios
manually or automatically. Figure 3 shows the Simio Facility layout of the model.
The Manufacturing Plat Orders consist of orders for both Intermediate Manufacturing Product and
Finished Products. The three types of intermediate manufactured materials are – Green Bulk, Red Bulk,
and Blue Bulk, which are mixed in available Mixing machines and later directed to the available Tanks.
The order for finished good materials makes use of the stored manufactured material as described before.
Finished products are first directed to the Filler machines and are later packed in the Packing Machines.
The model also takes into consideration the requirements of Raw materials, such as bottles and labels, that
are needed during the production process. It is imperative that the workers and manufacturing material
are available at each step/machine to ensure smooth execution. The Processing time and Setup time for
Manufacturing material and finished products on each machine is modeled as Triangular Distribution. This
accounts for variation resulting due to machine downtime, errors in setup process, or other uncontrollable
factors which cause variations in the manufacturing process. The model is enhanced by implementing
Simio’s custom dashboard features that display material, order details, and dispatch lists for use by operators.
The model is simulated to generate a 30-day Operation Planning and Production Schedule based on the
input orders, their attributes, resource constraints, and time availability. The following parameters were
captured for the experimental analysis - Total Cost of operations, Tardiness Cost – corresponding to the
late orders, Average Lateness, Number of late orders, and average time in the system. The Figure 2 show
the block diagram of Process flow once the order is generated.
The efficiency of any production unit is highly dependent on the resource utilization and production
schedule in execution. Following that, the model was tested for multiple scenarios by altering the availability
of operators and dispatching rules for different processes. The production unit is designed to work in 3
shifts – with each shift requiring operators according to the production schedule. A set for four dispatching
Dehghanimohammadabadi, Thiesing, and Belsare
Figure 2: The process flow of simulation model.
Least Setup Time (LST) - Setup time required to initiate an operation on a machine. Each type
of intermediate manufacturing material and finished material has different setup time on different
Earliest Due Date (EDD) - Decided based on Due dates mentioned in the Manufacturing Order
Largest Priority Value (LPV) - Priority as assigned in the Manufacturing Order table (range - 1,2
or 3)
Largest Attribute Value (LAV) - Decided based on degree of Lateness.
can be applied in different permutations to the Mixing and Packing operations. If there exists a tie for
sequence of Dispatching rule for an order then the one with least EDD is selected. The Mixing Dispatching
Rule represents the sequence of dispatching rules for processing Mixing and Tank Fill operations required
for intermediate manufacturing orders. Similarly, the Packing Dispatching Rule - represents the sequence of
dispatching rules for Filling and Packing operations. Hence, the number of operators in the shift and sequence
of dispatching rules were selected as Decision variables to analyze it’s effect on Total manufacturing Cost
and Tardiness Cost. The SODT framework used NSGA-II as a Multi-Objective Metaheuristic algorithm
to evaluate the effect of change in the number of workers in each shift and sequence of dispatching rules
on Total Manufacturing Cost and Tardiness Cost.
The NSGA-II initiates with a population that represents different number of operators in the shift and
sequence of dispatching rules. The algorithm then investigates the trade-off between different objectives.
When one objective cannot be improved without the worsening of another objective, we are on what is
Dehghanimohammadabadi, Thiesing, and Belsare
Figure 3: The simulation model.
known as the ‘Pareto front’. Chromosome length represents the total number of decision variables. Which
are- the sequence/order of the 4 mentioned dispatching rule for Mixing operation and Packing operation
individually plus number of workers required in the three shifts. Making the chromosome length as 11. The
initial parent population was set to 100 and from which 4 generations or children population were produced.
The maximum and minimum number of children per generation were kept to 80 to 50 respectively. The
crossover breeding was carried out using tournament selection method based on fitness scores with the
crossover probability of 0.8. The mutation probability was kept being 0.09. And the algorithm terminated
once the all the members selected for Pareto front were evaluated. Being ran on CPU and not GPUs the
computation time for total execution of said model was forty minutes. This can be further optimized using
GPUs and testing hyper-parameters of the model.
Figure 4: Pareto front graphs for the experiments performed. Objectives for experiment 1 were Total Cost
and Tardiness Cost. Objectives for experiment 2 were Total Cost and Time in the system.
The Figure 4 shows the Pareto Front graphs that were obtained from two different experimental analyses.
The objectives selected for the first experiment were Total Cost and Tardiness Cost. Upon analyzing, 3
Dehghanimohammadabadi, Thiesing, and Belsare
Pareto front solutions were obtained. Upon looking Table 1, we can observe that there is a lesser deviation
in Tardiness Cost among the Pareto Solutions but the Total cost has a considerable deviation. Achieving a
small deviation in the tardiness cost can be attributed to the smart permutations of dispatching rules and
balancing the number of operators in each shift. Given the due dates of the orders remain the same, these
Pareto solutions could help the decision-makers to perform a trade-off between operator availability and
Total Cost by merely altering the dispatching logic of the system.
Another experiment was run to test the relation between Total Cost and entity’s Time in the system.
Upon analyzing, 27 Pareto solutions were found as seen in the Figure 4. As seen in Table 2 the given
model presents a considerable variation in near-optimum solution to perform a trade-off between the given
objectives. Table 2 gives a glimpse of Pareto solutions obtained for this experiment. The values displayed
in the table represent the extremes and center point of the Pareto graph. Following are the abbreviations
used for the dispatching rules - Largest Attribute Value - LAV, Earliest Due Date - EDD, Least Setup Time
- LST, Largest Priority Value - LPV
Table 1: Results for Experiment 1.
Solution Cost Tardiness
Number Operator at
Shift-1, 2, 3
Sequence of Packing
Dispatching Rule
Sequence of Mixing
Dispatching Rule
Solution 1 4652.34 1609.42 2, 4, 4 LAV, LST, EDD, LPV LPV, LST, LAV, EDD
Solution 2 4663.60 1607.24 4, 3, 5 LST, EDD, LPV, LAV LPV, LST, LAV, EDD
Solution 3 4683.40 1606.93 5, 5, 5 LAV, LST, EDD, LPV LPV, LST, LAV, EDD
Table 2: Results for Experiment 2.
Solution Cost Time in
Number Operator at
Shift-1, 2, 3
Sequence of Packing
Dispatching Rule
Sequence of Mixing
Dispatching Rule
Solution 1 4084.81 119.12 1, 2, 3 LAV, EDD, LST, LPV EDD, LST, LAV, LPV
Solution 2 4339.74 88.12 1, 2, 1 EDD, LAV, LST, LPV LPV, LAV, LST, EDD
Solution 3 4646.05 57.61 3, 4, 3 LST, LPV, LAV, EDD LST, LPV, LAV, EDD
This paper demonstrates successful implementation of the Simulation-optimization framework for the Digital
Twin model of a Beverage Manufacturing Plant. The DT model is more than just a virtual representation
as it integrates real-time data tables to build the simulation environment. The SODT framework is not
only capable of harnessing the power of simulation engine but also capable of simultaneously optimizing
the search space. The proposed approach can help maximize the utilization and efficiency of the plant by
continually optimizing the DT model. With the help of Multi-Objective pareto front obtained from the
SODT framework, decision makers can have a clearer picture of the production schedule in execution.
The proposed SODT framework has the ability of rapidly adapt to the changes in orders, perform iterative
optimization and analyze multiple scenarios to provide essential feedback.
The proposed SODT is a promising approach that can be extended to various future works. The
experimentation example solely focused on the dispatching and labor tables. In fact, the used simulation
model is developed using multiple input tables and each of these tables can subject to optimization. One
interesting future extension could be to optimize order schedules and improve their release time to the
manufacturing floor. Another example would be analyzing the impact of layout changes on the model.
Since all resources are listed in a table, their coordinates can be easily changed to make new layouts.
This experimentation can be done without manual intervention or sophisticated layout design software
packages. The SODT model can develop multiple layouts based on the user expectation and evaluate them
Dehghanimohammadabadi, Thiesing, and Belsare
instantaneously. Another important advantage of this model is its capability to capture unexpected events
on the real-world system and provide immediate responses. Other analyses could include studying the
effect of machine failures, and new project/order arrivals using the SODT model. The proposed SODT is
very promising and leads to numerous future works.
Al-Ali, A., R. Gupta, T. Zaman Batool, T. Landolsi, F. Aloul, and A. Al Nabulsi. 2020. “Digital Twin
Conceptual Model within the Context of Internet of Things”. Future Internet 12(10):163.
Balderas, D., A. Ortiz, E. M´
endez, P. Ponce, and A. Molina. 2021. “Empowering Digital Twin for Industry 4.0
using metaheuristic optimization algorithms: case study PCB drilling optimization”. The International
Journal of Advanced Manufacturing Technology 113(5):1295–1306.
Ding, K., F. T. Chan, X. Zhang, G. Zhou, and F. Zhang. 2019. “Defining a digital twin-based cyber-
physical production system for autonomous manufacturing in smart shop floors”. International Journal
of Production Research 57(20):6315–6334.
Gyulai, D., J. Bergmann, A. Lengyel, B. K´
ar, and D. Czirk´
o. 2020. “Simulation-based Digital Twin of
a Complex Shop-Floor Logistics System”. In 2020 Winter Simulation Conference (WSC), 1849–1860.
He, B., and K.-J. Bai. 2020. “Digital twin-based sustainable intelligent manufacturing: a review”. Advances
in Manufacturing:1–21.
Kritzinger, W., M. Karner, G. Traar, J. Henjes, and W. Sihn. 2018. “Digital Twin in manufacturing: A
categorical literature review and classification”. IFAC-PapersOnLine 51(11):1016–1022.
Kunath, M., and H. Winkler. 2018. “Integrating the Digital Twin of the manufacturing system into a decision
support system for improving the order management process”. Procedia Cirp 72:225–231.
Leng, J., D. Yan, Q. Liu, H. Zhang, G. Zhao, L. Wei, D. Zhang, A. Yu, and X. Chen. 2019. “Digital
twin-driven joint optimisation of packing and storage assignment in large-scale automated high-rise
warehouse product-service system”. International Journal of Computer Integrated Manufacturing:1–18.
Liu, M., S. Fang, H. Dong, and C. Xu. 2020. “Review of digital twin about concepts, technologies, and
industrial applications”. Journal of Manufacturing Systems.
Liu, Z., W. Chen, C. Zhang, C. Yang, and Q. Cheng. 2021. “Intelligent scheduling of a feature-process-
machine tool supernetwork based on digital twin workshop”. Journal of Manufacturing Systems 58:157–
Park, K. T., S.-W. Jeon, and S. D. Noh. 2021. “Digital twin application with horizontal coordination for
reinforcement-learning-based production control in a re-entrant job shop”. International Journal of
Production Research:1–17.
Santos, R., J. Basto, S. G. Alcal´
a, E. Frazzon, and A. Azevedo. 2019. “Industrial IoT integrated with
simulation-A digital twin approach to support real-time decision making”. In Proceedings of the
International Conference on Industrial Engineering and Operations Management, July, 23–26.
Zhang, L., L. Zhou, and B. K. Horn. 2021. “Building a right digital twin with model engineering”. Journal
of Manufacturing Systems 59:151–164.
Zhang, M., F. Tao, and A. Nee. 2021. “Digital twin enhanced dynamic job-shop scheduling”. Journal of
Manufacturing Systems 58:146–156.
Zhang, Z., Z. Guan, Y. Gong, D. Luo, and L. Yue. 2020. “Improved multi-fidelity simulation-based
optimisation: application in a digital twin shop floor”. International Journal of Production Research:1–
Zheng, Y., S. Yang, and H. Cheng. 2019. “An application framework of digital twin and its case study”.
Journal of Ambient Intelligence and Humanized Computing 10(3):1141–1153.
Dehghanimohammadabadi, Thiesing, and Belsare
MOHAMMAD DEHGHANIMOHAMMADABADI is Assistant Teaching Professor of Mechanical and
Industrial engineering, Northeastern University, Boston, MA, USA. His research is mainly focused on
developing and generalizing simulation and optimization frameworks in different disciplines. This article
is part of his initiative to develop a framework to integrate a DES Simulation Optimization with Digital
Twin. His e-mail address is
RENEE THIESING is a Vice President of Products at Simio LLC. She is responsible for the design,
development, support and training of the Simio products. She has been working with Simio for over a
decade and has provided consulting services and training to Simio customers in the fields of healthcare,
logistics, supply chain, manufacturing, food production and aerospace. She has a BS and MS of Industrial
Engineering from the University Of Michigan and Auburn University. Her e-mail is
SAHIL BELSARE is a graduate student pursuing MS in Industrial Engineering at Northeastern University,
Boston, MA, USA. His research includes simulation combined with various optimization techniques like
Metaheuristics and Reinforcement Learning. Currently, he is working on developing simulation-optimization
frameworks for industrial optimization. His e-mail address is
Full-text available
In a re-entrant job shop (RJS), an entity can visit the same resource type multiple times; this is called re-entrancy, which occurs frequently in actual industries. Re-entrancy causes an NP-hard problem and is dominated by heuristics-based production control. The stochastic arrivals due to re-entrancy require the design of an appropriate dispatching rule. Reinforcement learning (RL) is an efficient technique for establishing robust dispatching rules; however, only a few cases that coordinate RL-based production control with a digital twin (DT) have been reported. This study proposes a novel production control model that applies a DT and horizontal coordination with RL-based production control. The requirements for dispatching in the RJS and coordination between RL and the DT were defined. A suitable architectural framework, service composition, and systematic logic library schema were developed to exploit the advanced characteristics of the DT and improve the existing production control methods. This study is an early case of coordinating RL and DT, and the findings revealed that RL policy networks should be imported in the creation procedures rather than being synchro-nised to the DT. The results should be a valuable reference for research on other types of RL-based production control with regard to horizontal coordination. ARTICLE HISTORY
Full-text available
The fourth industrial revolution, Industry 4.0, has been characterized by novel concepts introduction in manufacturing systems that enable smart factories with vertically and horizontally communication to improve their performance. Many virtual systems allow to predict foul conditions, save energy, study special cases, and so on, yet they need to implement new digital tools that allow developing manufacturing process in a better manner. As a result, Digital-Twin platforms are a good alternative since they are virtual models that could receive online and offline data. Thus, programmed algorithms can be evaluated to know the performance of the manufacturing process. These virtualizations and interconnections between elements of the manufacturing process become important components with an increasing role in dealing with supply, production times, and delivery chains as they run in parallel and find optimal performance before implementing these conditions into the real system. This study focuses on the use of a Digital-Twin that integrates a metaheuristic optimization and a direct Simulink model for printed circuit boards (PCB) design and processing focused on the drilling process. The results show that metaheuristic optimization can be integrated into the Digital-Twin concept as part of the production system into the drilling process. In the first part, it shows that depending on the penalization the optimization focuses on the lower path and forgets on changing the tools, yet as the penalization raises it focuses on finishing drilling with one tool before changing. Second, it is important where on the PCB it starts the drilling, with less time depending on each plaque. Third, it can be observed that using optimization can triple the amount of PCBs that can be manufactured. Finally, on an 8-hr run the Digital-Twin that didn’t use optimization can only work with three different designs, differently with optimization it can have 7-8 changes in the PCB design.
Full-text available
As the Internet of Things (IoT) is gaining ground and becoming increasingly popular in smart city applications such as smart energy, smart buildings, smart factories, smart transportation, smart farming, and smart healthcare, the digital twin concept is evolving as complementary to its counter physical part. While an object is on the move, its operational and surrounding environmental parameters are collected by an edge computing device for local decision. A virtual replica of such object (digital twin) is based in the cloud computing platform and hosts the real-time physical object data, 2D and 3D models, historical data, and bill of materials (BOM) for further processing, analytics, and visualization. This paper proposes an end-to-end digital twin conceptual model that represents its complementary physical object from the ground to the cloud. The paper presents the proposed digital twin model’s multi-layers, namely, physical, communication, virtual space, data analytic and visualization, and application as well as the overlapping security layer. The hardware and software technologies that are used in building such a model will be explained in detail. A use case will be presented to show how the layers collect, exchange, and process the physical object data from the ground to the cloud.
Full-text available
Various kinds of engineering software and digitalized equipment are widely applied through the lifecycle of industrial products. As a result, massive data of different types are being produced. However, these data are hysteretic and isolated from each other, leading to low efficiency and low utilization of these valuable data. Simulation based on theoretical and static model has been a conventional and powerful tool for the verification, validation, and optimization of a system in its early planning stage, but no attention is paid to the simulation application during system run-time. With the development of new-generation information and digitalization technologies, more data can be collected, and it is time to find a way for the deep application of all these data. As a result, the concept of digital twin has aroused much concern and is developing rapidly. Dispute and discussions around concepts, paradigms, frameworks, applications, and technologies of digital twin are on the rise both in academic and industrial communities. After a complete search of several databases and careful selection according to the proposed criteria, 240 academic publications about digital twin are identified and classified. This paper conducts a comprehensive and in-depth review of these literatures to analyze digital twin from the perspective of concepts, technologies, and industrial applications. Research status, evolution of the concept, key enabling technologies of three aspects, and fifteen kinds of industrial applications in respective lifecycle phase are demonstrated in detail. Based on this, observations and future work recommendations for digital twin research are presented in the form of different lifecycle phases.
In recent years, the concept of digital twin (DT) is attracting more and more attention from researchers and engineers. But there is still no consensus on what a right DT is. On one hand, some common models are renamed as DTs. On the other hand, some DTs extremely pursue ‘the same’ as physical objects, which bring unnecessary complexities to them. In this paper, we try to answer two questions from the point of view of model engineering: how to define a right digital twin, and how to build a right digital twin. The concept and related technologies of model engineering are introduced. Some basic principles and a set of metrics for a right DT are given. An evolutionary concurrent modeling method for DT (ECoM4DT) is proposed not only inheriting the theory from classic M&S methods but also highlighting the characteristics of DT compared with traditional models to systemically guide the DT modeling process.
Modern manufacturing enterprises are shifting toward multi-variety and small-batch production. By optimizing scheduling, both transit and waiting times within the production process can be shortened. This study integrates the advantages of a digital twin and supernetwork to develop an intelligent scheduling method for workshops to rapidly and efficiently generate process plans. By establishing the supernetwork model of a feature-process-machine tool in the digital twin workshop, the centralized and classified management of multiple data types can be realized. A feature similarity matrix is used to cluster similar attribute data in the feature layer subnetwork to realize rapid correspondence of multi-source association information among feature-process-machine tools. Through similarity calculations of decomposed features and the mapping relationships of the supernetwork, production scheduling schemes can be rapidly and efficiently formulated. A virtual workshop is also used to simulate and optimize the scheduling scheme to realize intelligent workshop scheduling. Finally, the efficiency of the proposed intelligent scheduling strategy is verified by using a case study of an aeroengine gear production workshop.
Due to numerous uncertainties such as bad weather conditions, frequent changes in the schedules of vessels, breakdowns of equipment, port managers are aiming at providing adaptive and flexible strategic planning of their facilities, especially intermodal terminals. In this research, we investigate a two-stage optimization of intermodal terminals main parameters via using AnyLogic simulation platform. We have developed a set of hybrid simulation models to optimize the main parameters of intermodal terminals which are also called dry ports. To make an express evaluation of the preliminary implementation of dry ports, we have developed an agent-based system dynamics simulation model to achieve the stable state of the main parameters of intermodal terminals. To clarify the obtained averaged benefits of the main dry ports parameters while the port managers make key decisions on the investments into implementation of intermodal terminals, we have developed an agent-based discrete-event simulation model of a seaport – a dry port system. We show that the combination of the agent-based modeling with other simulation approaches simplifies the process of designing simulation models and increases their visibility. The developed set of models allows us to compute the balanced values of the parameters, while an effective operation of a seaport – intermodal terminal system is achieved. On the basis of the provided case study on one of the busiest ports in China, we prove the adequacy and validity of the developed simulation models. Due to the lack of systematic approach to optimization of the main parameters of intermodal terminals in logistic industry, our findings set herein could improve the decision-making process related to the selection of strategic facility planning in the field of intermodal terminals.