The management of in-mine ore transportation systems is a crucial challenge due to the constant development of underground mines. The transport network's stability and efficiency directly affect enterprise profits and growth. Managing key areas such as transportation asset maintenance, object control, and ore extraction planning is essential to ensuring high transportation standards. For larger, multi-site enterprises, transportation supervision is divided among many managers. To address short-term production planning and management support during unplanned events, a decision support system is proposed in this article to meet the expectations of both lower and higher-level managers. The proposed concept is based on the simulation deployed from real-time operational data from mine, in which an agent is placed and taught through reinforcement learning. The simulation as well as agent possible actions have a block structure, where each means of ore transport is a separate add-on. In addition, it is assumed that the solution will be further improved by the implementation of various algorithms found in the literature (e.g. early fault detection algorithms). When created and implemented, such a system will bring improvements in multiple levels of operation, such as the safety of employees, overall machine health, cost reduction, energy efficiency, etc. As a result of this work, a conception of an intelligent decision-support system is presented, along with the specification of areas of its influence and a description of tasks and algorithms proposed to be incorporated within the system itself.