Home > Optimization Technologies > Evolutionary Algorithms

Optimization can be defined as the process of finding the best solution to a problem that has many possible solutions.

Evolutionary algorithm (EA) optimization is a directed search technique that evaluates hundreds of thousands of possible solutions as it converges on the best solution alternatives.

To apply EA optimization to a water system, an EA routine is linked to a hydraulic simulation model set up for the appropriate steady-state or extended period simulation (EPS) scenarios.

The EA search then sorts through different combinations of pipe, tank, pump and valve improvements, and/or operational set points and pumping schedules. The search objective is to find the best mix of decisions to meet the utility's design and performance criteria at the least cost.

The EA optimization approach contrasts with traditional simulation analysis where the designer uses trial-and-error and engineering judgment to evaluate a handful of trial solutions. Although a hydraulically feasible solution can usually be found, it is very likely the cost of the simulation solution is much higher than it needs to be.

Unlike traditional optimization techniques, such as linear programming (LP), dynamic programming (DP), and non-linear programming (NLP), EAs don't require well-behaved linear, differentiable, convex or continuous functions - so there is no need to simplify the problem. Also, EAs search from a collection or "population" of points. This allows the EA technique to in effect "climb many peaks" at the same time, resulting in a smaller likelihood of missing the global optimum.

One could be assured of finding the global optimum solution to a water system problem only if the problem was small enough to allow every possible solution to be evaluated, i.e., complete enumeration. Real-world water problems are rarely small. If a utility wants to look at its year 2020 improvements or its main replacement needs, it is not uncommon to consider options for 100 - 300 pipes. If each pipe option has say 10 allowable choices, then the size of the total solution space is 10¹⁰⁰ - 10³⁰⁰. (Note that these figures far exceed the number of atoms in the universe, estimated at 10⁷⁵).

Genetic Algorithm Optimization Genetic Algorithms are the most well known of all the Evolutionary Algorithms. Read More...

Ant Colony Optimization Ant colony optimization replicate the way ants search for food... Read More...

Particle Swarm Optimization Swarm Intelligence is modelled on the way in which a swarm converges to an object. Read More...