A Soft-Computing Approach for Decision Systems in Ring Imaging Cherenkov Detectors

In this paper the ability of the genetic approach has been investigated to recognize pattern produced by Ring Imaging Cherenkov, an high energy physics detector. The RICH patterns taken into account have been collected during an ALICE dedicated test beam at CERN. ALICE is A Large Ion Collider Experirnent, optirnized for the study of heavy-ion collisions. The RICH detector is designed with a proxirnity focusing geornetry for partic1e identification. When a charged partic1e pass through the radiator, the active rnediurn of the RICH detector, an annular flow of the uItravioIet photons is emitted which is subsequent1y proxirnity focused, onto the photon detection pIane. The sensitive pIane is coupled with the acquisition systern. The process for partic1e identification starts with Cherenkov light emission. It occurs at different rnornenta above a suitabIe threshoId for different partic1e rnasses. The rneasured effect is the production of an annular light region on the photon detection pIane. The effect has been tuned to produce circular regions with different radii in order to support the distinction among different c1asses of charged partic1es. Genetic Algorithrns, GAs, are taken into account to evaluate the feasibility of thern to support the decision systern in ALICE experirnent to identify c1asses of partic1es RICH-produced. At the purpose, the genetic process life cyc1estarts with a randorn population of individuals each one with a genetic characteristics described by chrornosorne. Then the evolution continues appIying the selection, rnutation and crossover rnechanisrn for the search process toward individuals of higher fitness. The detailed study to map the c1assification problern in natural evolution terrns together with experirnental resuIts will be presented.