All About Engineering Optimization

By Author: Vijay K Shetty
Total Articles: 2542
Comment this article

In order to achieve the desired results of designing in engineering and its applications optimization techniques are often used. This is called Engineering Optimization. The other name of Engineering Optimization is Design Optimization. The topics with which it deals include shape optimization, inverse optimization, processing planning, structural design, topographical optimization, product designs and many others. Under the section of Structural design, comes the design of welded beams and pressure vessels etc.
Topological optimization includes airfoil among others.

There are in general, three methods or techniques used to solve the problems of such optimization. These are evolutionary algorithms also known as genetic algorithms, more popular in its short form GA; traditional deterministic algorithms and metaheuristic algorithms.

In order to solve simple problems, the common traditional algorithms like hill climbing and the Hooke-Jeeves pattern search finds wide application. For problems that are more complex, the evolutionary strategies and algorithms are more widely used. The most recent among these are however the metaheuristic algorithms that are very promising as well. Among the metaheuristic algorithms are genetic algorithms, simulated annealing, harmony search, particle swarm optimization, differential evolution and many more of them.

The “simple problems” mentioned before are those problems, which include a sole minimum or a single minimum. Therefore, due to this fact, the minimum that is found is also the global minimum. On the other hand, more complex problems have more than a single minimum; they are ones that have local minima of many numbers. It may not be possible in this case, to find the global minimum using the gradient technique, although it may be capable of finding a local minimum.

Therefore, it is best to use the metaheuristic algorithms to solve problems for those methods, using a large number of initial search points as for example in genetic algorithms. Metaheuristic algorithms like particle swarm optimization along with the others are more competent in finding out the global minimum.

Particle swarm optimization is a technique that solves the problem by trying to improve a solution of candidate related to a given quality measure using the iterative method. This technique does not make use of the problem's gradient. Therefore, it is a good solution to irregular, altering and noisy optimization problems. Simulated annealing however, finds out a good approximate of the global minimum of a particular function in a big search space. Differential evolution is used for functions that are multidimensional and real-valued. It is a very similar technique to particle swarm optimization. Harmony search is a process, which gets its inspiration from the improvisation techniques of musicians. It is a phenomenon, which mimics algorithm.

Vijay K Shetty, Platinum Level Author.
Get more information on Engineering Optimization, Automation Software