Computational Management Science Бизнес, управление и учёт (прочее)
A bi-objective mixed-integer optimization framework for industrial project portfolio selection under sustainability and uncertainty
Аннотация
Industrial organizations increasingly face the challenge of selecting and scheduling investment projects while balancing economic performance, sustainability efficiency, and limited operational resources. This paper presents a bi-objective mixed-integer optimization framework for industrial project portfolio selection that jointly maximizes discounted net present value and sustainability efficiency under multi-period resource and budget constraints. Sustainability efficiency — defined here as a relative performance score derived from data envelopment analysis that benchmarks each project’s environmental and social outputs against its resource inputs — is used as the second objective alongside net present value, while uncertainty in project parameters is represented using deterministic, fuzzy, and Z-number formulations to reflect varying levels of information reliability. The resulting model incorporates realistic industrial constraints — including resource limits, project dependencies, and optional execution interruption — and uses an ε-constraint method to generate a Pareto frontier of portfolios balancing economic and sustainability objectives. A numerical case study comprising 15 candidate projects demonstrates that Z-number representations yield portfolio NPV estimates approximately 21% lower than deterministic equivalents, reflecting the reliability-discounting effect of explicitly modeling both parameter imprecision and expert confidence. Allowing optional project interruption increases portfolio NPV by up to 46% relative to continuous-execution solutions while yielding improved Pareto-efficient portfolios through greater scheduling flexibility. Three projects appear in every optimal portfolio regardless of the uncertainty regime applied, identifying robust core investments. These results illustrate how uncertainty representation and execution flexibility jointly shape portfolio composition and the economic–sustainability trade-off.