“Simulation is the next best thing to observing a real system in operation since it allows to study the situation even though we are unable to experiment directly with the real system” 
There are several reasons why a system would not allow a direct exploration of its behaviour under certain circumstances; for instance, the system may not exist yet or changing the system settings might prove too time consuming or expensive with respect to the foreseen gains. The complexity of a system is generally a constraint when mathematical methods, such as algebra, calculus or probability theory, are used to model it. On the contrary, there is almost no limitation to the complexity and types of system that can be simulated . This is proven by the breadth of areas where simulation has been successfully applied; these areas range from manufacturing to ecology and environmental issues, from business to biosciences and healthcare.
Discrete Event Simulation
Discrete event simulation (DES) is a very common simulation approach in the manufacturing field. With this technique, the evolution of a system over time is modelled so that changes of its state variables are allowed at discrete points in time. At these points, some events occur and, as a consequence, the state of the system may change. For instance, if a production line that includes machines and buffers is modelled, the completion of a processing operation on an item at a machine can be considered an event that will change the system status. The item will leave the machine and will be moved to the downstream buffer, provided that the buffer has available capacity.
Figure 1 Discrete Event Simulation flow
DES is capable of answering key operational questions relating to throughput, resource allocation, utilization, supply and demand. Typical applications of DES in manufacturing include:
- Design and evaluation of new manufacturing processes – assembly line balancing, capacity planning, facility location, etc.
- Performance improvement of existing manufacturing processes – feasibility studies on the introduction of new technologies (i.e. automated material handling system), quality management, logistics management, etc.
- Definition of optimum operational policies - Process improvement, start-up problems, equipment problems, performance measurements, etc.
- Complex evaluation function for an algorithm (or engine) to support production planning, inventory control, and scheduling.
The development of a simulation model generally follows a few fundamental steps –.a
- Problem definition - A comprehensive knowledge of the project scope is fundamental to ensure a simulation project success since the purpose of the analysis has relevant implications on model building and experimental design.
- Resource requirements
- Identification of simplifying assumptions – it is important that only the system elements and constraints that are relevant to the analysis to be performed are included in the model. Simplifying assumptions will reduce the model complexity without impacting the model significance.
- Conceptual model development - the systems characteristics to be measured should be clearly defined and the model should be conceptualised in a manner that an accurate prediction of these measures is catered for.
- Input data processing – this includes the collection of input data and their preparation into a format that the model can handle.
- Model development – Different simulation packages and languages can be used to develop a DES model.
- Verification and validation - Different verification and validation techniques are available and should be chosen based on the model characteristics .
- Results analysis and documentation – experimental plans can be executed and the simulation model can be used to support decision making processes.
When a valid and accurate simulation model is built, it can certainly be used to gain insights into the system behaviour and investigate the impact that different variables have on the system’s performance. The possibility of predicting the impact that system settings have on its performance make simulation a fundamental tool to support management decision processes and optimise a system. The versatility of simulation models and their capability of generating outputs relative to a wide input domain make them suitable for being integrated with optimisation techniques. From an optimisation viewpoint, a simulation model can be thought of as a function of an unknown form that transforms input parameters into output performance measures . This interpretation of simulation models as valid alternatives to mathematical functions has enabled their use with response surfaces and metamodels . A response surface is a numerical representation of the unknown function; on the contrary, a metamodel is an algebraic approximation of the function itself. Once the function to be optimised is obtained, either in numerical or algebraic form, classical optimisation techniques, such as random search, stochastic approximation, gradient-based approaches, response surface methodology, etc., can be applied and an optimal solution can be found. More recently, meta-heuristics and Artificial Intelligence (AI) are used in conjunction with DES; this enables a hybrid simulation-based optimisation approach that can be used to optimise complex systems whose improvement processes will otherwise rely on tacit knowledge of industrial practitioners.
Figure 2 Typical simulation-based optimisation flow
Finally, Integration of DES with Virtual Reality (VR) is proving to have a great value as VR and DES combinations allow both a modeler and the user to better understand a subject area. The combination of DES with VR particularly suits the virtual (or smart) manufacturing context as the live interaction with and manipulation of virtual manufacturing scenarios (which is enabled through 3D virtual systems) in real time with DES is a realistic goal of VR DES. VR can also assist simulation users in the model validation process. A user wearing a VR device will be able to view the running DES model superimposed onto the production line, allowing the user to interact with the model in real time, making decisions based on the outcome of what if experiments .
Figure 3 Example of virtual factory in Witness VR environment 
Simulation is a powerful management tool that can be used to better understand manufacturing system dynamics and explore tactical and operational improvement opportunities, with almost null investment costs.
 R. E. Shannon, “Introduction to the art and science of simulation,” in Proceedings of the 30th conference on Winter simulation, 1998, pp. 7–14.
 R. E. Shannon, “Introduction to simulation,” in Winter Simulation Conference, 1992, pp. 65–73.
 W. D. Kelton and A. M. Law, Simulation modeling and analysis. McGraw Hill Boston, 2000.
 R. G. Sargent, “Verification and validation of simulation models,” in Winter Simulation Conference, 2005, pp. 157–169.
 J. April, F. Glover, J. P. Kelly, and M. Laguna, “Practical introduction to simulation optimization,” in Simulation Conference, 2003. Proceedings of the 2003 Winter, 2003, vol. 1, pp. 71–78.
IEEE Trans. Human-Machine Syst., vol. 46, no. 6, pp. 882–894, 2016.