What Are Computer Simulation Games?

Computer simulation is a method that can be used to help business managers make decisions under uncertain conditions. The business manager must choose an action plan from a number of plans without fully understanding how the event occurred and its impact. If there is an ambiguous event, what will be the result, this is also uncertain. In some cases, the final impact on the results themselves is also uncertain. So there are two methods: decision tree method and computer simulation.

Establish a mathematical model or description model of the research object and embody and test it on a computer. The research objects include various types of systems. Their models refer to the general description of the system with the help of concepts, variables, rules, logical relationships, mathematical expressions, graphics, and tables. This mathematical model or description model is converted into a corresponding computer-executable program, and after inputting data such as system parameters, initial state, and environmental conditions, calculations can be performed on the computer to obtain results and provide various intuitive forms According to the analysis of the results, the relevant parameters or part of the structure of the system model can be changed, and the calculation can be performed again.

Chinese name: Computer simulation

Applied discipline: computer science

Introduction to Computer Simulation

Computer simulation, also known as computer simulation, refers to a computer program used to simulate an abstract model of a particular system. ^[1]

Computer simulation history

The development of computer simulation is inseparable from the rapid development of the computer itself. Its first large-scale development was an important part of the famous Manhattan plan. In World War II, in order to simulate the process of a nuclear explosion, Monte Carlo method was used to simulate with 12 hardball models. Computer simulation was originally used as a supplement to other aspects of research, but when people discovered its importance, it was widely used as a separate subject. ^[1]

Computer simulation

Usually divided into the following categories:

Discrete simulation
Analog simulation
Probe-based simulation
Simulation of stochastic processes or deterministic models ^[1]

The advantages and disadvantages of computer simulation

On the issue of applying computer simulation for risk analysis, it should be pointed out finally: this method requires obtaining the probability distribution of many variables such as investment expenditure, unit sales volume, product price, input factor price, asset life, and so on. Programming costs and computer operating costs. As a result, holistic simulations are generally not applicable (except for decisions on large, expensive plans such as expanding large plants or producing new products). In these exceptional cases, when companies are deciding whether to implement a large-scale plan that will cost millions of dollars, computer simulations can help evaluate the pros and cons of each alternative. ^[1]

Computer simulation development process

When people design and construct complex systems, or when studying the long evolutionary processes in nature and human society and things that are not easy to repeat, if you test the research object itself, you need to pay expensive considerations from time, manpower, and material resources. The cost is even impossible. Therefore, it is necessary to make a model for various experiments. ^[1]

In order to simulate a system, the system to be studied must first be identified or expressed. A mathematical model can be used to determine a system more conveniently and fully reflect the existing knowledge of the system or hypotheses that need to be verified, but it lacks intuitiveness and is not convenient for experiments. On the basis of the mathematical model, a physical model can be further made, which reflects the relevant properties of the real system that people require, but it need not be completely consistent with the real system in form and scale. Experiments with physical models are relatively intuitive and reliable, but they are still not economical and convenient. ^[1]

After the emergence of programmable digital computers, because of its strong mathematical operations and data processing capabilities, mathematical models can be compiled into computer programs to provide new and universal test methods. Computers can also be used to simulate operations-related activities, for example, to simulate the steps taken and the final outcome of the two parties participating in the competition. Its application field soon expanded to various types of systems, from large-scale systems to small-scale systems. The mathematical description of these systems is often very complicated, and it is very difficult to give complete analytical solutions or accurate numerical solutions. Computer simulation through repeated experiments helps people understand the performance of the system, test the expected hypotheses, perform system analysis, design, prediction or evaluation, and can also provide a fairly realistic environment to train and train personnel. Computer simulation has become a powerful tool in many fields such as engineering research, natural science research, economic and social research, teaching and training activities, military research, and organizational management. ^[1]

Basic methods of computer simulation

Computer simulation usually goes through many steps from problem formation to final model validation. Form problems and clarify the purpose and requirements of the simulation. Collect and process system-related data as much as possible. Form a mathematical model, find out the various components that make up the system, and describe the relevant variables (generally including input variables, state variables, and output variables) or parameters of their states at each moment; determine the rules of interaction and influence between each component , That is, the functional relationship between these description variables. When selecting parameters and variables, you must also consider whether they can be identified or solved, and whether the model is finally suitable for testing based on data from real systems. Determine or estimate the parameters in the model based on the collected data, and select the initial state of the model. Design the logic or information flow chart until the computer program is compiled. Program verification, checking the consistency between the program and the mathematical model, and the rationality of the input amount. Carry out a simulation test and execute a program on a computer for a given input. Result data analysis, collecting and sorting test results and making explanations. If necessary, the input amount or part of the model structure can be changed and the experiment can be repeated. The model confirms the consistency between the results obtained by the model and the performance data of the real system. This is a key issue related to the effectiveness of computer simulation. It depends on the level of testing the real system itself, whether it can obtain sufficient observation data and the criteria for judging consistency. The level of model effectiveness can be divided into: reproducible, that is, the model can reproduce the performance of the real system; predictive, that is, the model can effectively predict the future performance of the real system; constitutive, that is, the model can reflect the interior of the real system Structure. Because the system itself changes with time or has randomness, the comparison of real system data and model test results often requires the use of time series analysis methods or statistical analysis methods. ^[1]

Computer simulation of discrete-time models

The time in the discrete-time model is represented as a sequence of integers (representing an integer multiple of a certain time unit), and only the state changes of the system at these moments are considered. A typical simulation program of this model includes the following steps: Set the initial value of the simulation time T to t 0. Set the initial value of the state variable. After giving the value of the current simulation time input variable, according to the state transition function in the model, determine the value of the state variable T = t + h at the next moment. Then determine the value of the output variable at that moment based on the output function in the model. Advance the simulation time T by a unit time h . Check whether the simulation time T reaches the predetermined termination time. Stop if it has reached; otherwise go to step . ^[1]

Computer simulation of discrete event models

In the discrete event model, the state change of the system only occurs at discrete moments, which is called a discrete event. Taking the queuing system as an example, the basic steps and methods for establishing such a simulation model are: Determine all relevant "entities" and their attributes contained in the system, all "events" that change the state of the system and their antecedents and consequences. Entities are components of the system, and the attributes of each entity are represented by numerical values representing their properties, which constitute the state of the system. The most basic entities in the queuing system are a certain number of "service stations" and "customers" that require service. Their attributes are "service rate" and "customer" service priority and arrival service system. Moment of waiting. Basic "events" include: new entities enter the system or existing entities leave the system, entity attributes change, and schedules change. Determine the method to simulate the passage of time. If the time is divided at equal intervals, and the system examines whether or not events occur at these moments, it is called a fixed time interval method; if the length of each time passes is based on the time at which the next event occurs, it is called a variable time interval method or " Next Event "method. Because the occurrence of events in the system is often random and obeys a certain probability distribution, it is necessary to generate random numbers of these distributions on the computer. In order to flexibly and effectively record the state of the system, schedule events, accumulate relevant performance data and form reports, save and automatically manage future event files, it is very appropriate to use database technology in program design. ^[1]

Computer simulation continuous system simulation

A system whose state changes continuously with time is called a continuous system, and the rate of state change satisfies a certain differential equation. Establishing a corresponding simulation model on a computer depends on effective numerical methods for solving differential equations and compiled into standard subroutines so that various equation orders, coefficients, initial value conditions, or boundary value conditions can be used. Simulations of systems that include feedback and control are typical examples of this type. ^[1]

Computer simulation language

Assembly language and general programming languages (such as FORTRAN, ALGOL, etc.) can be used when programming the simulation program. Various analog languages are also available. Computer simulation language is a high-level programming language for describing system models. It provides modules that represent many basic units, components, and scheduling operations in the system model. Users can use it to easily determine the basic structure of the model, as long as some additional programs are added, a simulation program can be compiled.

The simulation language is generally established on the basis of other general-purpose programming languages. It requires its own compiler to pre-compile, convert the simulation language program into a general-purpose programming language program, and then compile it again to convert it into an executable program on the computer. Simulation language can reduce the user's program work, but it also inevitably brings some restrictions, which consume more memory and computing time. ^[1]

Simulation languages can be divided into discrete event simulation languages (such as GPSS and its various modifications, SIMSCRIPT, GASD, CSL, SIMULA, etc.) and continuous system simulation languages (such as DARE, ACSL, CSS, CSSL, etc.). There are special simulation languages for various application areas.

Computer simulation is closely related to the development of computer hardware and software technology. In order to facilitate the establishment of the model and the validity test of the model, people try to make the simulation model have a certain degree of similarity in time and space with the real system. In the simulation process, I hope that I can easily change the parameters and even the structure of the model, and can output data and charts at any time through keyboard commands. Therefore, computer simulation requires computers to have strong parallel processing capabilities, high computing speed, human-computer interaction capabilities, and easy-to-use simulation languages. ^[1]

Computer simulation application

The scale of computer simulation can be either macro or micro. On a macro scale, experimental databases can be used to predict the process flow, operating conditions and system properties, calculate the mechanical and processing properties of materials, and are generally used in chemical process simulation, machinery manufacturing and other fields. At the micro-scale, the structure and properties of micro-particles play an important role, and are generally used for reaction mechanism research and macro-property simulation. ^[1]

Computer simulation examples

To illustrate this method, let us study the construction of a textile factory. The construction cost of the plant has not been accurately calculated, and is estimated to be about 150 million US dollars. If no difficulties occur during construction, the cost could be as low as $ 125 million. But it is also possible that investment expenditures as high as $ 225 million will occur due to various unexpected events such as strikes, unexpected price increases in raw materials, and technical problems. ^[1]

The new plant will be able to operate for many years, and its product sales income depends on the growth of population and residents' income in the region, the degree of competition in the same industry, the research and development of synthetic fibers, and the import quota of foreign textiles. Its operating costs will depend on production efficiency, the trend of rising raw materials and wage levels, and so on. Since sales revenue and operating costs are uncertain factors, annual profits are also uncertain. ^[1]

Assuming a probability distribution can be calculated for each of the major cost and income factors, a computer program can be built to simulate possible events. The computer actually takes any value from each relevant distribution, combines it with other values selected from other distributions, and provides an estimate of the amount of profit and the net present value of the investment, or profit rate. This specific amount of profit and profit rate is only suitable for the combination of specific values selected for this experiment. The computer continues to select the values of the other groups, and it is possible to calculate other profit amounts and profit margins for hundreds of experiments. The number of calculations of different profit margins is counted and saved. After the computer finishes running, a frequency distribution can be drawn according to the number of occurrences of different profit margins. ^[1]