Metaheuristic & Heuristic Optimization in Python

Explore how genetic algorithms optimize problems by evolving a population of solutions through selection, crossover, and mutation to improve fitness across generations and expand the solution space.

Explore genetic algorithms, including population, chromosome, genes, and fitness function, and how selection, crossover, mutation, and elitism drive evolution to optimize problems; compare generational and steady-state models and termination conditions.

Explore how genetic algorithms simulate evolving defense strategies under dynamic warfare, optimizing mission success, resource utilization, and casualty reduction while revealing adaptable, resilient military planning insights.

Explore a math model for defense optimization: define actions, resources, scenarios, and parameters; use binary decision variables and a weighted objective to maximize combined effectiveness under resource and feasibility constraints.

Explore coding with DEAP to optimize defense strategies via a genetic algorithm, defining resources, actions, and scenarios; build a fitness function, create individuals, apply crossover and mutation, and evolve solutions.

Analyze DEAP's results by tracking generations, fitness metrics, and hall of fame, revealing how selection, crossover, and mutation improve the best solutions over time.

Explore the A* search algorithm, an efficient way to find the shortest path between two nodes by combining Dijkstra's algorithm and greedy best-first search with f = g + h.

Explore the star search algorithm, an efficient route finder for graphs that blends Dijkstra's algorithm and greedy best-first search using f = g + h to identify the shortest path.

Explore a SaaS optimization model that navigates a five by five city grid of intersections and routes, adapting edge travel times to peak and non-peak traffic for time-efficient deliveries.

Model route as graph with intersections as vertices and roads as edges whose weights vary by time, using a decision variable and constraints to minimize time start to destination.

Implement a weighted A-star search with the Hipc library’s priority queue on a 5x5 city grid, using Manhattan distance and accounting for peak-hour travel times to compute an optimal route.

Observe how the A-star algorithm navigates a city grid from 0,0 to 4,4 using the Manhattan distance heuristic, accounting for 8 a.m. peak hour congestions and dynamic travel costs.

Explore particle swarm optimization, a swarm-based method using personal and global bests to guide velocity and position for search; apply to wind turbine blade design, length, twist angle, and material.

Define a model with decision variables—blade length, twist angle, and material density—bounded by limits. Maximize blade efficiency under four constraints and a budget defined by four parameters and cost coefficients.

Use particle swarm optimization in Python to maximize wind turbine blade efficiency by optimizing length, twist, and density within bounds, yielding blade length 50 m, twist 0, density 8000.

Explore the hill climbing algorithm by iteratively evaluating neighboring points from a random start to solve problems, reaching a local optimum, applying it to drone delivery routes and vehicle routing.

Explore hill climbing as a simple heuristic for optimization, starting from a random point, evaluating neighboring solutions, and iterating toward better moves until reaching a local optimum with no improvements.

Explore drone delivery optimization by minimizing total travel distance from the depot through delivery points and back, modeled as a drone vehicle routing problem to improve efficiency and sustainability.

Define all nodes, including the depot and deliveries, with distance parameters and binary variables to minimize total distance while visiting each location once and returning to the depot.

Apply a Python hill climbing approach to optimize a drone routing problem using a distance matrix and depot, starting from a random route and reversing segments to minimize distance.

Apply hill climbing to a drone delivery route and obtain an optimized path represented as 042310 with a total distance of 16, starting and ending at the depot.

Explore multi-objective genetic algorithms and Pareto front optimization to balance green spaces, residential, commercial, and industrial areas, and minimize noise in a sustainable urban layout through evolutionary methods.

Examine the mathematical model allocating land to zones with a fixed area to maximize parks, environmental quality, and commercial and industrial zones, while minimizing industrial noise to residents.

Examine a multi-objective genetic algorithm to optimize urban planning by maximizing green space and economic viability while minimizing noise, using Python, and later refine with Gurobi.

Gurobi finds an optimal objective of 1000 in about 0.018 seconds, using four decision variables with 100 and 300 unit zone allocations.

Explore simulated annealing, a metaheuristic inspired by metallurgical annealing, to navigate complex solution spaces with temperature and the Metropolis criterion. Apply it to routing and delivery optimization, traveling salesman problem.

Implement a simulated annealing algorithm in Python to optimize routes from a distance matrix, computing total distance, reversing subsequences, and iterating with temperature-based acceptance to escape local minima.

Explore how simulated annealing identifies an optimal route from a distance matrix, revealing a sequence of locations that minimizes total distance and demonstrates the value of optimal distance for routing.

Explore the harmony search algorithm, an optimization technique inspired by musical harmony, using harmony memory and improvisation to balance exploitation and exploration while minimizing transportation costs in supply chains.

The lecture models the harmony search optimization with decision variables for transport quantities and supplier activation, detailing costs, demand, capacity constraints, and a cost-minimizing objective.

This lecture explains a Python harmony search approach to optimize transportation costs and supplier activations, using harmony memory, hmcr, par, and bw to improve feasible solutions.

Learn how the harmony search algorithm yields the best solution vector and best cost for a supply chain optimization, with the vector representing transport quantities and supplier activation statuses.

Explore variable neighborhood search in Python, plugging in objective functions and neighborhood structures, with optional local search to tackle the traveling salesman problem using a flexible VNS framework.

Explore artificial fish swarm algorithm for optimizing the Rastrigin function with the Circuit of Python library, building a Python model and evaluating navigation of local minima toward the global minimum.

Explore the artificial fish swarm algorithm inspired by fish schooling, including swarming, following, and guiding behaviors. Learn how scikit-opt provides a Python implementation for practical optimization.

Investigate the Rastrigin function with the artificial fish swarm algorithm, using Python and scikit-opt to locate the global minimum amid local minima.

Minimize a function with the artificial fish swarm algorithm, where each fish is a potential parameter set navigating a bounded, multi-dimensional search space under constraints toward the objective function.

Implement and run an artificial fish swarm algorithm to minimize the two-dimensional Rastrigin function in Python, tuning parameters like fish_num, maxiter, visual, and step, and reporting best_x and best_A.

Review the outcomes of the artificial fish swarm algorithm on the decision function, showing the optimal solution near 0.003, the minimum value around -0.00549, and progress toward the global minimum.