In this paper an extension of task graph for real-life optimization problems is presented. Such an extension allows to divide tasks on several groups which helps in finding the optimal way of tasks execution. The groups of tasks determine if and which tasks can be executed by one or more resources. It also determines the potential types of resources able to execute the tasks. What is worth to underline this increases the number of situations that can be described using graphs even if standard task graph could not deal with them. Therefore extended task graph can be used in every situation in which task graphs are used and in many more. Proposed extension can help in many optimization problems by making it easier and cheaper to solve.