EVOLUTIONARY ALGORITHMS in architecture design
- 1. EVOLUTIONARY ALGORITHMS UNIT - III V.Raghuram Professor/Architect
- 2. Introduction to Evolutionary Algorithms •Definition: Evolutionary Algorithms (EAs) are optimization techniques inspired by the process of natural selection. •Key Concepts: •Population-based search. •Genetic operations: selection, crossover, mutation. •Fitness function: measures how good a solution is.
- 3. What is Algorithm? An algorithm is a specific, step-by-step procedure or a set of instructions designed to solve a problem or perform a specific task. It outlines the logical steps needed to transform input data into desired output, much like a recipe guides you through the steps of cooking. In computer science, algorithms are essential for writing programs and solving computational problems
- 4. Evolutionary Algorithms in Architecture Evolutionary Algorithms are computer-based problem-solving methods that are inspired by the natural process of evolution – just like how animals and plants evolve over generations to survive better. 👉 In EA, designs evolve by trying different versions, keeping the best ones, and combining them to create new, better versions – like nature does with genes.
- 5. 📌 How it works: Start with a group of random designs (called a "population"). Evaluate each design (based on some goal – like least material, best ventilation, most sunlight, etc.) Select the best designs. Combine them (crossover) and mutate them slightly to make new designs. Repeat the cycle to keep improving!
- 6. What is Evolutionary Art? In art, evolutionary algorithms are used to create patterns, paintings, or forms by evolving visual ideas over time. 🎨 Example: Imagine a computer generating a bunch of abstract paintings. A designer picks the most beautiful one. The computer then combines and modifies those to make new ones. This cycle continues, and after many generations, we get amazing artworks created through evolution.
- 7. 🔍 Real Example: Karl Sims is a digital media artist and visual effects software developer. His interactive works have been exhibited worldwide including at the Pompidou Center, Ars Electronica, DeCordova Museum, Boston Museum of Science, and the National Museum of Mathematics. He founded GenArts, Inc. which created special effects software tools for the motion picture industry, and he also held positions at Thinking Machines Corporation, Optomystic, and Whitney/Demos Productions. Karl studied computer graphics at the MIT Media Lab, and Life Sciences as an undergraduate at MIT. He is the recipient of various awards including two Ars Electronica Golden Nicas, a MacArthur "Genius" Grant, and an Emmy Award.
- 9. Evolutionary Algorithms (EAs)** Evolutionary Algorithms (EAs) in architecture refer to the application of evolutionary principles —such as selection, mutation, and crossover—used in computational design processes to generate innovative architectural solutions.
- 10. 1. Selection: Selection is like choosing the best options from a set of possibilities. Just like in nature, where the fittest organisms are more likely to survive and reproduce, in evolutionary algorithms, the "best" solutions are chosen to continue to the next generation. Example: Imagine you have several designs for a building. The best design, based on criteria like space efficiency or aesthetic appeal, is selected to form the foundation of the next design.
- 11. “It is not the strongest of the species that survives, nor the most intelligent, but the one most responsive to change.” by Charles Darwin
- 12. Example from Nature Giraffes live in areas with tall trees, so those with long necks can reach the higher leaves and have a better chance of surviving. There were two types of giraffes in the past: long-necked and short-necked. Long-necked giraffes had a survival advantage because they could access more food and win fights, leading them to breed more. Short-necked giraffes, on the other hand, struggled to survive and reproduce. Over time, the long-necked giraffes became more common, and short-necked giraffes went extinct. The mutation for long necks provided a survival advantage in nature.
- 14. 2.Mutation Mutation is a random change made to a design. It’s like introducing a small new variation to a solution. This randomness helps explore new possibilities and prevents the algorithm from getting stuck on one solution. Example: In architecture, imagine a building’s shape is slightly changed—maybe the roof’s angle is adjusted or the windows are reshaped. This change might bring unexpected improvements.
- 15. Example for Mutation This mutation in architecture— shifting from traditional to sustainable design—has become a survival advantage in a world facing climate change and resource scarcity. Just like in nature, over time, this type of design will likely become the norm, as the benefits in terms of environmental preservation, cost savings, and occupant health continue to prove its value.
- 16. 3. Crossover: Crossover is like mixing the best parts of two solutions. It’s inspired by how genes combine during reproduction in nature. The idea is to combine different "good" parts of existing solutions to create a potentially better solution. Example: Let’s say you have two building designs: one with a fantastic layout and the other with a great facade. Crossover would combine the layout of one with the facade of the other to create a new design that might be even better.
- 17. Result
- 18. How They Work Together: Selection picks the best designs. Crossover combines parts of these good designs. Mutation randomly changes some aspects of the new designs to explore new possibilities. Through these processes, the algorithm keeps improving solutions (like architectural designs) over multiple generations, just like how nature evolves over time!
- 19. Chromosome In genetic algorithms, a chromosome represents a potential solution to a problem, often encoded as a string of genes
- 20. Genetic Algorithm The Genetic Algorithm helps you find the best solution by evolving your choices over time using a combination of selection, crossover, and mutation. It’s like trial and error, but smarter, and it gets better over time.
- 21. STRUCTURE
- 22. Example - Choosing the Best Vegetarian Dish to Cook Problem: Choosing the Best Vegetarian Dish to Cook for the WeekendLet’s say you’re trying to choose which vegetarian dish to cook this weekend. You have afew options in mind, but you want to make the best decision, something exciting
- 23. Step 1 ( Different Solutions) Representation of Solutions (Chromosomes)You have a list of vegetarian dishes you are considering: Dish 1: Vegetable Biryani Dish 2: Paneer Butter Masala Dish 3: Stuffed Paratha Dish 4: Vegetable Stir-fry Each of these dishes is like a chromosome in the Genetic Algorithm world. Each dish is apossible solution to your weekend cooking decision.
- 24. Step 2: Initial Population Initial Population Now, you start with your initial population. You could randomly pick a couple of dishes, orjust pick all of them to begin with. For this example, let’s say you randomly pick 4 options.Your initial options could be: Dish 1: Vegetable Biryani Dish 2: Paneer Butter Masala Dish 3: Stuffed Paratha Dish 4: Vegetable Stir-fry
- 25. Step 3: Fitness Function Now, you need a way to rate or measure how good each dish is for the weekend. Let’s makethe fitness function based on a few factors like: 1.How quick and easy is it to cook? (Easier dishes are better!) 2.How much you like the dish? (Personal taste!) 3.How many ingredients do you have at home? (Less shopping is better!)
- 26. Fitness score
- 27. You might give each dish a score out of 10 for each of these factors: Vegetable Biryani: 8 (easy to cook, you like it, but needs some ingredients) Paneer Butter Masala: 6 (takes more time, you like it but need extra ingredients) Stuffed Paratha: 7.3 (easy to cook, you like it, but it needs specific ingredients) Vegetable Stir-fry: 8.7 (quick and easy, you like it, and you have most ingredients)
- 28. Step-4 Selection you need to select the best dishes based on their fitness scores. You will choose thedishes with the highest scores to move to the next step. From the table, Vegetable Stir-fry has the highest score (8.7), followed by VegetableBiryani (8). These will be selected to be the parents for creating new dishes.
- 29. Step 5: Crossover (Mating) you combine features of the best dishes (mating) to create new dishes. For example: Parent 1: Vegetable Stir-fry (9, 8, 9) Parent 2: Vegetable Biryani (7, 9, 7) You can combine the best features of both: Take the quick and easy part from Vegetable Stir-fry (9/10). Take the taste from Vegetable Biryani (9/10). Take the ingredients availability from Vegetable Stir-fry (9/10). Now, your new offspring dish could be a quick and tasty stir-fried Biryani! New Dish (Offspring): Quick Veg Biryani Stir-fry (combining both!)**
- 30. Step 6: Mutation add mutation. Mutation is like making a small random change to try and improvethe result. For example, you might decide to try adding a new ingredient to the Quick VegBiryani Stir-fry, like tofu or a new spice that you’ve never tried before. This is the mutationstep.The new dish could be: Quick Veg Biryani Stir-fry with TofuNow, this dish might have a slightly better taste or texture that could make it even better thanthe original!
- 31. Step 7: Repeat continue the process of evaluating, selecting, crossing over, and mutating your dishes until you find the one that you’re most happy with. You might try this over a few weekendsuntil you find the perfect weekend dish.
- 32. Summary 1. Initial Population: Choose a list of possible dishes to cook. 2. Fitness Function: Rate each dish based on factors like how easy it is to cook, howmuch you like it, and how many ingredients you already have. 3. Selection: Choose the dishes with the highest fitness scores (best options). 4. Crossover: Combine the best features of two dishes to create a new dish. 5. Mutation: Make small changes to try new things (like adding a new ingredient). 6. Repeat: Continue testing and adjusting until you find the perfect dish. By repeating this process, you evolve your choices and eventually pick the best vegetarian dish for your weekend
- 33. Key Idea: Genetic Algorithm helps you find the best solution by evolving your choices over time using a combination of selection, crossover, and mutation. It’s like trial and error, but smarter, and it gets better over time.