Optimal Investment in the Development of Oil and Gas Field

TL;DR

This paper addresses the complex problem of optimally investing in oil and gas fields with capacity constraints, proposing an approximate algorithm that performs close to optimal in practice.

Contribution

The paper introduces an approximate algorithm for investment optimization with pipeline capacity constraints, providing qualitative analysis and numerical validation.

Findings

The algorithm yields solutions close to the optimal in numerical experiments.

The problem is NP-hard, but the proposed method offers practical approximation.

Qualitative results support the effectiveness of the algorithm under capacity restrictions.

Abstract

Let an oil and gas field consists of clusters in each of which an investor can launch at most one project. During the implementation of a particular project, all characteristics are known, including annual production volumes, necessary investment volumes, and profit. The total amount of investments that the investor spends on developing the field during the entire planning period we know. It is required to determine which projects to implement in each cluster so that, within the total amount of investments, the profit for the entire planning period is maximum. The problem under consideration is NP-hard. However, it is solved by dynamic programming with pseudopolynomial time complexity. Nevertheless, in practice, there are additional constraints that do not allow solving the problem with acceptable accuracy at a reasonable time. Such restrictions, in particular, are annual production volumes. In this paper, we considered only the upper constraints that are dictated by the pipeline capacity. For the investment optimization problem with such additional restrictions, we obtain qualitative results, propose an approximate algorithm, and investigate its properties. Based on the results of a numerical experiment, we conclude that the developed algorithm builds a solution close (in terms of the objective function) to the optimal one.