Artificial Intelligence: Optimization Algorithms in Python

Explore a flights dataset in python by loading a text file into colab, modeling flights with an origin-destination dictionary storing departure, arrival, and price, and preparing optimization for cheapest options.

Represent the flight scheduling problem in Python with a list-based encoding that captures city names, origins, destinations, and times, enabling optimization algorithms to find best schedules while minimizing waiting time.

Create a get_minutes function to convert hours and minutes into total minutes, using the time library for parsing, enabling airport waiting-time calculations with tests like 6:13 and 0:00.

Craft the fitness function, heart of an optimization algorithm, by combining time and price into a cost to minimize, accounting for constraints like waiting time and early or late flights.

Implement a basic random search to generate flight schedules, evaluate them with a fitness function, and select the best solution within a defined domain; study how trials affect results.

Compare hill climb with random search, showing 500 iterations yield similar results to 10,000 calls, about 20 times faster. Demonstrate that using a predefined initial solution further reduces time.

Learn the theory of simulated annealing, inspired by metal cooling. Start from a random solution, lower temperature, and change one value at a time, occasionally accepting worse moves.

Learn how to implement simulated annealing in python, defining the domain and cost (fitness) function, configuring temperature and cooling, and generating neighbor solutions to improve results.

Implement a crossover function in Python for a genetic algorithm, using a random cut point to combine left genes from one solution with right genes from another, and test.

Implement a step-by-step function to allocate bedrooms to each person, using a vacancies list and bed indexing to print the solution for a constrained optimization in python.

Compare optimization algorithms by multiple trials and averages; the genetic algorithm is best, while hill climb and simulate underperform random search on this small dataset.

Represent intelligent problems in Python, implement flight-schedule optimization with random search, hill climb, simulated annealing, and genetic algorithms, compare results, and apply to minimization and maximization tasks using Rose Library.