Multi-Objective Optimization with Python Bootcamp A-Z

Explore the Google Colab environment, create and run notebooks, save and export code as ipynb or py, connect to Google Drive or GitHub, and troubleshoot with Stack Overflow.

Discover numpy basics by creating 1d and 2d arrays, inspecting shape, size, and ndim, and performing indexing, slicing, and element-wise operations, with a note on matrix multiplication versus arrays.

Explore how genetic algorithms mimic natural selection through initialization, selection, crossover, and mutation, using giraffes and peppered moths to illustrate evolutionary processes.

Form a bi-objective optimization with three variables x1, x2, x3 and two constraints. Note F1 and F2 conflict and constraints must be converted to ≤ form for the PMO library.

Install the pymoo library in Colab or your preferred IDE using pip, with per-session and one-time setup instructions for Windows and Mac.

initialize the nsga two multi-objective genetic algorithm for a two-objective problem using pymol, with float sampling, sbx crossover, and pgm mutation, and terminate after 100 generations.

Learn to configure and run a multi-objective optimization with an NSGA-II style algorithm, converting the second objective to minimization, and interpreting Pareto, non-dominated, and epsilon indicators.

Discover the Pareto front, a set of non-dominated solutions in multi-objective optimization, revealing trade-offs between objectives and how to approximate it through decision making and exploration.

Learn how to tackle multi-criteria decision making (MCDM) by evaluating multiple conflicting objectives with quantitative methods like compromise programming and pseudo weights using the PMML library.

Learn compromise programming as a quantitative decision method. It covers the augmented scalarization function using lambda weights and the optimization of decision variables and weights to minimize the ESF.

Learn how pseudo weights, a simple multi-criteria decision making method, assign and normalize weights and objectives, compute a weighted sum, rank alternatives, and select the optimal one.

Calculate ideal and nadir points from the objective vectors. Normalize the objective vectors to the [0,1] scale and visualize them with Matplotlib, preparing for compromise programming or pseudo weights method.

Explore decision making with the pseudo weights method and compare its results to compromise programming in a multi-criteria decision making (MCDM) framework, using weights and normalized objective vectors.

Solve a three-objective maximization with F1, F2, F3 via NSGA-II, bound x1 and x2, then visualize in 3D and normalize with ideal and nadir using compromise programming or pseudo weights.