Evolutionary Algorithm

evolutionary algorithms mimic the biological process of evolution in nature and can very quickly search through a huge number of possible solutions in a large problem space they can be used to optimise a population of solutions; in other words, they can look at more than one solution in a single iteration this is critical because it allows for the simultaneous exploration and optimisation of different solutions in a short period of time some key characteristic of the evolutionary algorithm are that it is non deterministic (random) like bioligical evolution, but far faster highly iterative and based on route permutations and inheritance the evaluator the calculator within the engine assesses the performance of each route based on user defined criteria these feed into various "objectives" such as cost, penalty, visual impact or proximity to existing infrastructure the best routes are ranked by their objective value(s) and then proceed to the next iteration of the optimisation the optimization stops after a set number of iterations or when solutions stop continuing to improve the objective is not to find one single optimal design but to quickly find multiple solutions that balance costs and impacts, as configured by the user infrastructure planners and designers can then assess the trade offs between different solutions using their own expertise, decide which designs to take forward and work out where they need to commission site surveys to gather more detailed, accurate data critically, evolutionary algorithms open up the possibility of tracing back candidate results to understand what was calculated both before and after them, why a solution was taken forward, and why it might have been rejected from further assessment it should be noted that, with the rapid development of the artificial intelligence space, legitimate concerns have been raised regarding data privacy regarding ai models these concerns have remained in focus while making development decisions at continuum industries; we ensure the core algorithm within optioneer is not trained on, and does not learn from, any customer data customers can ensure a common approach to assessing project features and constraints without compromising the data security of the organisation and their approach to project development we maintain, manage and constantly update our solution and have a customer success team to guide customers through projects and train them in using optioneer to achieve project aims