Systems Simulation and Processes

Systems Simulation and Processes

• Simulation originates from the Latin word "simulare," meaning "to pretend"
• Powerful technique that safely conducts experiments on a system model
• The quality of the system model significantly impacts the simulations
• Manipulation of a system model operating in time/space allows perception of relations and interactions not easily apparent otherwise
• Simulation helps in understanding problems where errors or faults are unacceptable
• It involves computerizing a developed model and running it over time to study defined interactions, it is generally iterative
• Functions as an operational research tool optimizing system performance, improving processes, planning facilities, bridging the gap between analysis and intuition

Advantages of Simulation

• Offers a deeper system understanding
• Provides insight into resource performance
• Allows trying out multiple alternatives
• Saves time and money
• Generates virtual environments
• Improves analysis quality

Predicting Complex Systems

• It's difficult due to high system complexity and insufficient core theory
• Issues arise from uncertainties, dynamic interactions, and complex variable interdependencies

Types of Simulation

• Physical Simulation: Experiments with a physical prototype using real, scaled input to compare outputs with a real system
• Numerical Simulation: Uses mathematical models, is popular, easy to modify, and proceeds sequentially in a time-based manner
• Digital (Computer) Simulation: Uses digital computers to simulate equations, has clarity, and features a high degree of automation

Characteristics of a Simulation

• Efficiency
  • Higher mission availability
  • Increased operational system availability
  • Transportation avoidance
  • Reduced or eliminated expendable costs
  • Less procurement and operational costs
• Effectiveness
  • Improves proficiency or performance
  • Forms fewer complex activities
  • Provides neutral and obstruction services
  • Enables greater observation, assessment, and analysis capability
• RIsk Reduction
  • Safety of stakeholders
  • Less environmental impact
  • Machinery and equipment handling

Simulation Techniques

• Monte Carlo (Risk Analysis) Simulation
  • Generates random sample data based on known distribution
  • Applies to risk quantitative analysis and decision-making
  • Used when estimations are necessary and uncertain

Monte Carlo Simulation Characteristics

• Generates random output samples
• Requires knowing the input distribution
• Requires knowing the results while experimenting

Practical Application of Monte Carlo Simulation

• Consider estimating construction project time
  • Use expert knowledge to determine absolute max and min times -Project has 3 parts done sequentially, total time is the sum of the parts

Forecasting Models in Construction Projects

• Basic forecasting model
  • Task Job 1: 5 months
  • Task Job 2: 4 months
  • Task Job 3: 5 months
  • Total is 14 months

Realistic Modelling

• Using a range of possible values helps do this
• When models are based on ranges, the output will also be a range
• Using minimum and maximum estimates helps estimate the total minimum and maximum time for the project

Monte Carlo Estimates

• Task Job 1: 4 months (minimum), 5 months (most likely), and 7 months (maximum)
• Task Job 2: 3 months (minimum), 4 months (most likely), and 6 months (maximum)
• Task Job 3: 4 months (minimum), 5 months (most likely), and 6 months (maximum)
• Totaling 11 months (minimum), 14 months (most likely), and 19 months(maximum)

Analysing Risk

• Creating a Monte Carlo simulation model helps to analyse risk
• Randomly generate job values and calculate the total completion time when running a simulation many times
  • Run simulation 500 times
• Counting how many times the model gives a particular result helps calculate likelihood
  • Useful to know how many times the simulation gave a result of less than or equal to a number of months

Completion Likelihood

• Simulation results on completion likelihood
• 12 months(0.2%)
• 13 months (6.2%)
• 14 months (34.2%)
• 15 months (78.8%)
• 16 months (96.4%)
• 17 months (99.8%)
• 18 months (100%)
• Original expected case for the "Most Likely" scenario was 14 months
• With simulations using random values, the total number of occurrences of 14 months or less totals to 34.2%

Simulation Results Analysis

• Only has a 34.2% chance that any individual trial will result in 14 months or less
• Has a 78.8% chance that the project will complete in 15 months
• Suggests that its unlikely that it will fall at the absolute minimum or maximum total time estimates

Project Risks Mitigation

• Simulation models demonstrates risks to project plan
• Different choices impact development
• More information at the start of the simulation will contribute to more accurate results and better planning

Agent-Based Simulation

• This focuses on individual active system components.
• Uses active entities such as people, vehicles, equipment etc
• Establishes connections between entities and their behaviors and also environmental varibales

Agent-Based Simulation Usefulness

• Use when:
  • interactions between entities are complex nonlinear and discontinuous
  • Space is vital and entities are not fixed
  • Population is heterogeneous and each individual is potentially different

Crowd Stampedes (Example)

• Panic induced stampedes lead to fatalities
• Triggered in life threatening situations or a rush for seats of seemingly no reason
• Fire escape and outflow of people are a good example

Agent-Based Simulation - Narrowed Problem

• Problem: The effect of placing a pillar before the exit slightly asymmetrically to the left and one meter away
• One might think the pillar will slow the out flow
• However agent-based simulation and real world indicate it regulates flow leading to fewer casualties

Discrete Event Simulation

• Focuses on the processes in a system at a medium abstraction level
• Used in manufacturing, logistics and healthcare
• Simulation technique codifies the behavior of a complex system as an ordered sequence of well defined events.

Characteristics of Discrete Event Simulation

• Comprises specific changes in a system state at a point in time
• Characterized by predetermined starting and ending points that are discrete events or instances in time
• Includes lists of discrete events that have occurred and incoming
• Must track the time since the process began
• Characterized by a graphical, statistical or tabular record of the function it is engaged with

Automated Teller Machine (Example)

• ATM example components
  • System states are the number of customers in line (Q(t)) and server status at time t (S(t))
  • Entities are customers and the server
  • Events are arrival (A), departure (D), stopping events (E)
  • Event Notices are specified in the form of an arrival event to occur at time t ((A, t)), or a customer departure at a time ((D, t)) or the simulation stop event at time Tsim ((E, Tsim))
  • Activities are inter-arrival time and service time Delay is time spent by each customer

System Dynamics Simulation

• Highly abstracted method of modeling and simulation
• Ignores fine system details and individual properties.
• Generates a general representation of acomplex system.
• Used for strategic modeling and simulation

Simulation Examples

• Piston providing force for a wheel to rotate
• The speed of the rotation, radius of the wheel and length of the rod vary