Schedulers
Companies daily need to optimize their products and processes, so optimization is becoming a keyword in today's design cycle. Problems with one or more than one objective originate naturally in several disciplines and their solution has been a challenge to engineers for a long time.
There are several sources of complexity in optimization, such as the computational difficulties in modeling the physics, the potentially high number of variables, or a high number of objectives and constraints. Each optimization technique is qualified by its search strategy that implies the robustness and the accuracy of the method.
There are thousands optimization methods in literature, each numerical method can solve a specific or more generic problem. A lot of "classical" optimization methods exist; these methods can be used providing that certain mathematical conditions are satisfied. Unfortunately, real world applications often include one or more difficulties which make these methods inapplicable.
modeFRONTIER is a very powerful tool for multiobjective optimization and it includes the most widely used methods. It contains both "classic" and metaheuristics methods for single and multiobjective optimizations. Metaheuristics methods are a new kind of methods that have been developed since 1980. These methods are inspired by analogies with physics, or with biology and they have the ability to solve difficult optimization problems in the best way possible; moreover, they have hard contributed to the renewal of the multi-objective optimization. This class of methods includes between the others: simulated annealing, genetic algorithms, evolutionary strategies.
All these metaheuristics are not mutually exclusive. It is often impossible to predict with certainty the efficiency of a method when it is applied to a problem. This statement is confirmed by the well-known "no-free-lunch theorem" (NFLT) developed by D. Wolpert and W. Macready.
For all these reasons, modeFRONTIER includes a wide range of possible algorithms that can be selected for solving different problems. At present, the methods available in modeFRONTIER are:
Classical gradient based algorithms use the "direction of improvement" information in order to achieve a fast and accurate convergence towards the optimal solution. Well-established mathematical bases guarantee their efficiency. More...
Genetic Algorithms (GA) use the analogy of natural evolution of species: natural selection and reproduction guide the evolution towards individuals that are better adaptated to the environment. In this way GA lead to optimal designs. More...
Simulated Annealing methods work on the basis of a thermodynamical analogy: the slow cooling of a heated system let it settle down in a configuration of minimum energy. Out of the metaphore it means that the optimum is reached. More...
Evolution Strategies (ES) are optimization technique based on ideas of adaptation and evolution. In this sense they are similar to GA, but here the main search procedure is a smart mutation operator. More...
RSM based algorithms. Response Surfaces Methodology (RSM) is a collection of mathematical techniques useful for modelling the output functions of interest. If RSM are incoropated within an optimization algorithm in an adaptive way, then the algorithm is speeded up considerably. More...
Other algorithms borrow good ideas from Nature: natural systems can suggest the best strategies to solve specific problems. More...