Applying artificial intelligence to the scheduling problem in the ERP system

The basic enterprise tasks are how to satisfy customer demands, how to fulfil due dates, low prices and undoubted quality. Today most of Croatian enterprises are not able to fulfil those obligations. Usually, enterprises have inappropriate planning, scheduling and launching models. The paper shows structure of Croatian Enterprise Resource Planning System (ERP) that is developed for single and batches production and place of artificial intelligence technologies in development of new generation of EPR system. Through the application of 3-tournament steady-state selection genetic algorithm scheduling process can be improved and plan variants can be achieved according to cost, time or both. At the beginning were only some planning techniques such as line charts, line of balance, organisational appliances. The computer era began the first computer programs so called informational islands. These programs solved only particular problems like problems connected with accounting, bookkeeping or employees.