Hi Chamanth, all depens on the mathematical form of your problem : continuous optimization, discrete optimization, mixed . Do you have a formulation of it?

Most of the heuristic optimization algorithms have more or less the same computational complexity and converge to the same solution. In this case, I will agree with what Richard is saying. It all depends upon the objective and the type of optimization that is required.

Richard Epenoy Since i am in initial stages of project i do not have the math ready and wanted to know what would be the best way to achieve the expected

(formulation for such problems) .

Syed A. Raza To be precise the main objective would be to optimize the residual load coming from the distribution grid by using the renewbles and at the same time find the best time for EV to charge so that the it does not overload the situation. I would like to know how these things are dealt in general .

You can try AIS : Artificial Immune System. It is based on machine learning, and used for optimization purposes in many fields.

Chamanth Kandalam , it will be better that you develop the maths first. Once it is developed and you know your objective function (whether to minimize or maximize it), application of an optimization algorithm should not be that difficult. you can refer to some of papers that on optimization in my profile.

Syed A. Raza can you to tell me which papers of yours is relevant to this ? unfortunately i am not able to figure out the right paper that involves optimization algorithms from the lot !

Thanks for the response though !!

Chamanth Kandalam , you can start with this:

Article A Differential Evolution Based Adaptive Neural Network Pitch...

Chamanth Kandalam I echo with previous comments.It all depends on your problem formulation, objective function and deciding min-max problem.Optimization problems are dealt in large scale in previous published fragmented literature.You can have a look at mathematical equations of previous papers.

I concur with the previous responses that having a mathematical formulation of your optimisation problem ready is pivotal for deciding on a suitable way to solve it.

In addition, I would also like to point you towards considering how important it is to you to find the global optimum to your problem or whether an approximate solution is sufficient.

While most heuristic (possibly nature-inspired or artificial intelligence) algorithms are fast and often provide solutions very close to the global optimum, there are no guarantees for that! You should always have some kind of naive benchmark strategy or a global solution if you seek to assess the performance of heuristic algorithms.

My impression is that for most continuous and mixed-integer linear problems as well as some convex non-linear problems, solvers like commercial Gurobi, CPLEX or open-source GLPK are the way to go. The speed is certainly also dependant on your number of variables and constraints and this is subject to your own preferences or the requirements of the problem at hand. The advantages of these solvers are you can always deduct a maximum distance to the global optimum.

Fabian Neumann The point is how accurate can some of these alogorithms be when computing the real time 'time series data' !!

Chamanth Kandalam, if I understand your point correctly, you are talking about applications in online (or near-real-time) optimisation. In this case heuristic solving algorithms may have their charm for they are usually fast; however if e.g. Gurobi converges within your time limits it will give you a globally optimal solution (within your specified tolerances). With the heuristics you can very often not be sure.

Math equations a must. Do you know the objective function for your problem? Try the FortranCalculus (FC) language that got its start by NASA's Apollo Space program and got us to the moon! It has dozen+ Solvers stored in a library for solving most equations.

FC-Compiler™ application is a (free) Calculus-level Compiler that simplifies Tweaking parameters in ones math model. FC is based on Automatic Differentiation that simplifies computer code to an absolute minimum; i.e., a mathematical model, constraints, and the objective (function) definition.

Most math Modeling, Simulations, & Optimizations can be easily solved with a Calculus-level language/compiler. A half-hour per problem should be enough time to code and execute most math problems. Solves Algebraic Equations through Ordinary Differential Equations, for more visit http://fortrancalculus.info/apps/fc-compiler.html . Equations may be implicit, non-linear, Boundary Value Problems (BVP), Initial Value Problems (IVP), etc. Download, install, & execute some demo problems. Simple and its Freeware! (PDE solvers are in the works!)

Their is a FC Casebook, http://fortranCalculus.info/textbooks/ , that shows 20+ industry problems with solutions. Just these examples should give you a head start on types of problems you may encounter in industry. The one stopping block for non-engineers or scientists is knowing a problems objective/goal(s). Without an objective, you can still learn how to use FC to solve problems. But, your solutions won't be optimal.

Links:

• https://www.researchgate.net/project/FC-Compiler-the-FortranCalculus-Compiler-Alpha-version
• https://www.researchgate.net/post/Automatic_Differentiation_what_is_it_and_where_is_it_used
• https://www.researchgate.net/post/Solve_YOUR_EQUATIONS_how_to