Abstract
The problem of distributing work volumes at different stages of construction or reconstruction of engineering networks is solved, taking into account the current state of communal utility systems. The problem is formulated considering the conditions of initial information uncertainty and the prospective development of the engineering network. The proposed mathematical model is a Boolean programming problem with variables in the form of interval numbers that characterize the assignment of design territory sections to a particular stage of construction or reconstruction of the engineering network. The problem is solved by Boolean programming methods taking into account the apparatus of interval arithmetic. Two variants of the algorithm for finding a compromise solution are proposed. Initially, numerical characteristics are calculated based on expert conclusions, which are associated with interval coefficients. At the next stage, the reduction task is solved by using two extreme objective functions, which limit the set of feasible compromise solutions. Based on the intermediate results, variables that take the same values are selected. The values of these variables are taken as final and are removed from the objective function of the initial model, which significantly reduces the number of variables in the model. Boolean programming methods are again applied to the new model with a smaller number of Boolean functions and interval coefficients of the objective function, which provides the most probable solution. If the problem has a solution, then it is close to the optimal one. If the problem does not have a solution, then fuzzy preference relation methods are used on a set of alternatives. If at the beginning of the modeling it is allowed to reduce the requirement that the site belongs to the corresponding stage of construction (reconstruction) of the engineering network, then Boolean programming problems are modified to partially discrete linear programming, which allows expanding the criteria for assigning final values to variables and removing them from the objective function of the initial model.