MATLAB (EX-506) - Unit 5 Short Notes

UNIT 5: ADVANCED MATLAB PROGRAMMING & APPLICATION DEVELOPMENT

5.1 Advanced Programming Constructs & Code Optimization

Goal: Write faster, memory-efficient MATLAB code.

[!TIP] Common Pitfall: Growing an array inside a loop (A = [A newElement]) is the #1 cause of slow MATLAB code. Always preallocate.

5.2 Graphical User Interface (GUI) Development

Goal: Build interactive applications with visual components.

Core UI Components (App Designer):

• UIAxes (for plotting)

• Button, EditField (text/numeric), DropDown, Slider, CheckBox, ListBox

Callbacks & Data Sharing:

• Callback function signature: function ButtonPushed(app, event)

• Share data via app properties (define in Code View → Properties section). <br> Example: app.SharedData = [1 2 3];

• Avoid global or handles in App Designer.

Deployment:

1. In App Designer: App Compiler tool.

2. Select "Standalone Application".

3. Specify main file (*.mlapp).

4. Runtime: Users need MATLAB Runtime (MCR) installed (free download).

5.3 Advanced Graphics & Visualization

Handle Graphics System:

MATLAB graphics are objects in a hierarchy:

Figure

└── Axes

    ├── Line (plot)

    ├── Scatter

    ├── Text

    └── Image

• Get handle: h = plot(x,y); (h is a Line object).

• Modify properties: h.LineWidth = 2; or set(h, 'LineWidth', 2);

Advanced Plot Types:

• Error bars: errorbar(x, y, err)

• Stem: stem(x,y)

• Scatter (varying): scatter(x,y,sz,c) (sz=size vector, c=color vector/colormap)

• Filled contour: contourf(X,Y,Z)

Subplots & Layouts:

• Modern: tiledlayout(m,n) + nexttile (better spacing/control than subplot).

  
  tiledlayout(2,2);
  
  nexttile; plot(x1);
  
  nexttile; plot(x2);
  
  

• Legacy: subplot(m,n,p)

Animations & Real-Time:

• Inside loop: drawnow limitrate (updates figure, limits rate to screen refresh) or drawnow.

• pause(0.01) controls speed.

• For high-speed: Update XData/YData of existing plot object instead of plot repeatedly.

5.4 Object-Oriented Programming (OOP) in MATLAB

Defining a Class:

classdef MyClass

    properties (Access = private)

        PrivateData

    end

    properties (Constant)

        PI = 3.14159;

    end

    properties (Dependent)

        ComputedProp % No storage; get/set methods define behavior.

    end

    methods

        function obj = MyClass(val) % Constructor

            obj.PrivateData = val;

        end

        function val = get.ComputedProp(obj)

            val = obj.PrivateData^2;

        end

    end

    methods (Static)

        function z = staticFunc(x)

            z = x*2;

        end

    end

end

Key Attributes:

• Properties: SetAccess (public/private), GetAccess, Constant, Dependent.

• Methods: Access (public/private), Static (no object needed), Hidden (not shown in docs).

• Handle vs. Value Classes:

  • Handle class: Inherit from handle. Objects are references. Changes affect all copies. Used for GUI components, graphics objects.

    
    classdef MyHandleClass < handle
    
    

  • Value class: Default. Objects are copies (like structs). Changes to copy don't affect original.

• Inheritance: classdef SubClass < SuperClass

• Packages: Folder name prefixed with + (e.g., +mypkg). Class inside: mypkg.MyClass.

Events & Listeners:

• Define event in events block: events; MyEvent; end

• Trigger: notify(obj, 'MyEvent')

• Listen: addlistener(obj, 'MyEvent', @callbackFunction)

5.5 Interfacing with Other Languages & Systems

5.6 Parallel and Distributed Computing

Parallel Computing Toolbox (PCT):

[!TIP] parfor vs for: Use parfor only if loop iterations are heavy and independent; overhead of parallelization can make it slower for small tasks.

5.7 Advanced Toolbox Applications (Selective Overview)

5.8 Testing, Debugging, and Documentation

Debugging:

• Breakpoints: Click margin or dbstop in file at line.

• Step: dbstep, dbstep in, dbstep out.

• Continue: dbcont, Quit: dbquit.

• dbstop if error (stops on error).

• dbstack (show call stack).

Unit Testing:

classdef MyTest < matlab.unittest.TestCase

    methods (Test)

        function testAddition(testCase)

            act = add(1,2);

            exp = 3;

            testCase.verifyEqual(act, exp);

        end

    end

end

Run: runtests('MyTest').

Documentation:

• publish('script.m', 'html') → Generates HTML with code, output, figures.

• Options: 'pdf', 'latex'.

• Comments: %% for sections. % TODO: recognized.

MATLAB Projects:

• matlab.project.createProject or via Project tab.

• Manages files, dependencies, source control (Git/SVN), tasks, shortcuts.

• Ensures reproducibility.

5.9 Deployment and Application Distribution

MATLAB Compiler (mcc):

• Command: mcc -m myapp.m (console app) or -W WinMain:myapp (Windows GUI).

• App Designer: Use App Compiler app → generates installer.

• Output: Executable + runtime folder (or web app archive).

Web Apps:

• Create App in App Designer.

• App Compiler → Web App.

• Deploy to MATLAB Web App Server (requires license).

• Users access via browser.

Toolbox Packaging:

• matlab.addons.toolbox.packageToolbox('mytoolbox.prj') → .mltbx file.

• Shareable, installable via Add-Ons → Install from File.

Runtime Dependencies:

• MATLAB Runtime (MCR) must be installed on target machine.

• Version must match MATLAB version used for compilation.

• Include any custom files (data, images) via Additional Files in compiler.

• Toolboxes: If code uses toolboxes, target must have corresponding toolbox runtime (some toolboxes require separate runtime installers).