An algorithm for planning work in a group of autonomous mobile robots performing the functions of eliminating the consequences of fires or emergency situations (ES) is proposed. Such a multi-agent robotic system (MRTS) should work effectively without the direct participation of an operator performing the functions of a decision maker (DM). The autonomy of the MRTS is due to the emergency conditions with foci of infection, dangerous for the stay of people. The paper proposes an approach and a corresponding algorithm for constructing a machine-learning model that solves the problem of optimal current planning for emergency response operations for MRTS, where an experienced decision-maker is the teacher.

Keywords: mobile robots, work area, job distribution, limited resources, linear programming problem, optimal planning, constraints

A method for indirect assessment of the effectiveness of the use of internal and external checks of the readiness of units of the State Fire Service to fulfill the tasks of responding to calls is proposed. The potential for efficiency gains is assessed by reducing various types of damage through improved response rates. The initial data for the assessment are statistical data on the temporal characteristics of the response of fire departments to calls, which are approximated by a gamma distribution. The estimation takes into account the regression dependences of damage on response indicators.

Keywords: control, damage, response time, gamma distribution, quantile, efficiency improvement potential