Skip to main content

9. Modelling

All Past Paper Questions: https://docs.google.com/document/d/1SwdogVb7VeJox38pnkpfD5TQInLHnXAgzq9NVoJiAwI/edit?usp=sharing

Models​

  • create models before building them in real life

  • we create models to run simulations

  • advantages

    • can test all aspects without damaging expensive hardware

    • can prodive results that are generally not measurable in some experiments

    • can allow you to see how a system might respond

      • can test and modify to avoid real mistakes

    • cheaper as no need to make different prototypes to test them out

    • safer when testing nuclear models

    • designing in computer saves time

    • gives accurate predictions of results

    • critical situations can be executed

      • without risk
      • more quickly

    • models can be slowed down to study more closely & carefully

    • we can ask many "what if?" questions

    • testing desings with computer is more safe

    • testing like this is cheap

    • can obtain results soon

      • even actions that take a long time in real life

    • re-designing computer models is cheaper

    • researches will have tools to simulate scenarios

      • that are hard to observe in operating reactors

  • disadvantages

    • expensive to buy modelling software

    • sometimes, data might be insufficient

    • simulator response is not exactly the same

      • as there are too many variables

    • requires fast computer system

      • with fast processor
      • and large amounts of storage

    • a quick event might take hours to imitate in a simulated environment

    • reduction of simulation, based on oversimplification of assumption

    • users might be given a false sense of security

    • researches need extra knowledge in setting the simulation up

      • need specialist knowledge (eg: of mechanical engineering)
      • most researches only specialize in one field
        • they need to learn another field
        • which costs money

    • cannot exactly re-create

    • not every variable is included in the model

    • leading to incaccurate results

    • specialist knowledge required to build the model

Spreadsheets​

  • For creating financial models, why use spreadsheets?
    • absolute and relative cell referencing
      • can only deals with data we want to
      • allows to change indidual items to see effect
    • cell protection
      • makes sure the cell doesnt get deleted by accident
      • for fixed data/costs, eg: overhead
    • user interface forms
      • easier to input data to model
    • macros
      • easier to make more complex functions
      • eg: comparing different data/costs simultaneously
    • automatic re-calculation
      • dont need to evaluvate formula everytime we change it
      • formulas can be re-calculated automatically
      • eg: changing indiviual data/costs
    • conditional formatting
      • allows to highlight values
      • that match given criteria
      • easily seeing what makes profit
    • goalseek
      • to know which variables to change
      • to reach goal
    • values can be changed to ask what if questions
    • how many good need to be sold
    • graphs
      • to show trends
      • and illustrate forecasts
      • eg: of which are likely to make profits
    • producing an effective model maybe time consuming
    • the might have repeated formulae
    • complex predictions may need third party software / modules
      • which are expensive to buy
    • data can be entered accurately using validation and verification

Simulations​

  • nuclear reactions

    • advantages
      • avoids safety problems, eg:
        • nuclear meltdowns
        • environmental problems
      • testing nuclear reactions using computer models cut costs
        • dont have to pay for meterials
        • eg: uranium
      • results can be obtained in a short period of time
        • when actual reactions might take a long time
      • re-designing the model is cheaper than re-creating a nuclear reaction
      • can be designed to allow extreme conditions
    • disadvantages
      • cant always re-create exact real world experiment
      • not all variables can be inlcuded
        • so, inaccurate results
      • costs money to hire computer experts to create the models

  • pilots using flight simulators

    • how they work
      • simulates difficult scenarios
      • need specialised equipment
        • costs a lot
        • eg: pilot seat, joystick, throttle, alititude gauges, etc...
      • need specialised software
        • has rules of model
        • that defines how the simultation works
        • includes database about information/map of all airpots
      • control panel and cockpit mimicks a real one
      • displays generated by computer
      • computer graphics produce display using
        • projectors
        • panoramic spherical mirrors
    • advantages
      • cheaper cuz no expenses for aircraft maintenance
        • no fuel costs,
        • no insurance costs, etc...
        • maintaining simulator is much cheaper
      • no pollution (air/noise)
      • doesnt put trainee, instructor & flight in danger
      • no cost to replace parts (incase of an accident)
      • emergency situations can be re-created quickly & easily
      • less time wasted travelling
      • low level of noise
        • makes communication between instructor & trainee easier
    • disadvantages
      • simulator response is not same an actual airplane
      • as there are too many variables
      • some pilots will be overconfident, and will mess up actual driving
      • pilots may be more stressed when riding real plane
      • length of training session is shorter
        • might be more tired in real flights

  • weather forecasting (predctions)

    • involves several weather forecasters
    • and days of calculations
    • today, one model can handle the task
    • models used to predict future patterns
    • computer models can help make relaibale predictions
    • can be used to predict future patterns of weather
    • more complex models make more accurate predictions
    • human forecasters still need to evaluvate the results
    • requires the use of a lot of data
    • must consider many variable to be effective & accurate
      • atmospheric pressure
      • humidity
      • rainfall
      • temperature
      • wind speed and direction
    • a lot of processing power is needed
      • so, super computers have to be used
    • models can compare current conditions with previous similar trends shown over the time.

  • teaching to drive

    • advantages
      • can teach without worrying about damage to expensive car
      • can provide variables hard to measure
        • eg: reaction time of driver
      • can allow you to see how a system might respond
        • before you design or modify it
        • avoiding mistakes
      • safer to use a model to simulate cards
        • safer for driver/instructor/passengers/pedestrians
      • can be slowed down to study behaviour
      • easier to simulate critical event
        • rather than waiting for it to happen
        • eg: different driving conditions
    • disadvantages
      • insufficient data to produce mathematical models
      • sometimes impossible to create accurate models
        • eg: effects of human behaviour
      • formulae and functions used may not be accurate
      • may give drivers a false sense of security when driving

  • safety drill about evacuvating people off a building

    • advantages
      • no need to stop working of large number of workers
      • leading to lack of productivity during that time
      • no risk of stempedes occuring
        • so, no risk of injuries
      • emergency procedures can be quickly produced / re-created
      • less time wasted training people to unwanted drills
      • no. of people in trial may not be representative
    • disadvantages
      • there may be too many variables
      • to exactly simulate the evacuation
      • high costs of employing experts to create the simulation
      • time taken to create simulation is higher
      • difficult to accurately predict
        • the exit doors chosen by each individual