Task Realization's Optimization with Earliness and Tardiness Penalties in Distributed Computation Systems

There are many service systems (especially in reservation systems, electronic commerce, in tasks synchronized directly with the Internet), where each task has to be executed in some fixed range of time. Exceeding the term is disadvantageous and causes additional penalties. Therefore, it is necessary to establish optimal sequence of tasks (which minimizes penalties) and its starting times. It amounts to some job scheduling problem with earliness and tardiness. Because usually tasks are received in the distributed system (in the web), so to solve the presented problem we propose a parallel coevolutionary algorithm based on the Lamarck evolution and the island model of migration.