Comparison of Guided Random Search Techniques to Minimize Objective Functions of Continuous Variables

D. HANSMANN; G. FIER; G. TERRANOVA

San Carlos de Bariloche

Congreso; TREFEMAC 2016 - 14° Congreso Regional de Física Estadística y Aplicaciones a la Materia Condensada; 2016; 2016

Universidad Nacional de Cuyo

Single-objective optimization (SOO) deals with the study of those kinds of problems in which one has tominimize or maximize functions of real or integer variables. In contrast to multi-objective optimization (MOO)(multi-criteria or multi-attribute optimization), SOO deals with the task of optimizing simultaneously one or morenon-conflicting objectives with respect to a set of constraints. In this context, in particular if the function isunimodal, one can choose among many good alternative algorithms, ranging from Newton?s method to SimplexMethods to Random Search Methods.In this work we compare velocity of convergence, precision, robustness and general performance of different ran-dom search optimization algorithms using a set of proper artificial landscapes (test functions) like Goldstein?Pricefunction, Ackley?s function, Beale?s function and others. The optimization algorithms include classical randomsearch techniques and particle swarm optimization (PSO) techniques, where the latter are based on dynamic ofswimming E. coli bacteria, known as run and tumble. In addition we test and compare several modifications andsimplifications of the E. coli dynamics that were used during the last years to model swimming and swarmingbacteria.