A metaheuristic algorithm for project selection and scheduling with due windows and limited inventory capacity

Abstract

Purpose – The purpose of this paper is to provide an effective solution for a complex planning problem encountered in heavy industry. The problem entails selecting a set of projects to produce from a larger set of solicited projects and simultaneously scheduling their production to maximize profit. Each project has a due window inside of which, if accepted, it must be shipped. Additionally, there is a limited inventory buffer where lots produced early are stored. Because scheduling affects which projects may be selected and vice-versa, this is a particularly difficult combinatorial optimization problem. Design/methodology/approach – The authors develop an algorithm based on the Metaheuristic for Randomized Priority Search (Meta-RaPS) as well as a greedy heuristic and an integer programming (IP) model. The authors then perform computational experiments on a large set of benchmark problems over a wide range of characteristics to compare the performance of each method in terms of solution quality and time required. Findings – The paper shows that this problem is very difficult to solve using IP, with even small instances unable to be solved optimally. The paper then shows that both proposed algorithms will in seconds often outperform IP by a large margin. Meta-RaPS is particularly robust, consistently producing the best or very near-best solutions. Practical implications – The Meta-RaPS algorithm developed enables companies facing this problem to achieve higher profits through improved decision making. Moreover, this algorithm is relatively easy to implement. Originality/value – This research provides an effective solution for a difficult combinatorial optimization problem encountered in heavy industry which has not been previously addressed in the literature.

Selección óptima del portafolio de proyectos utilizando metaheurísticas de población y trayectoria meta-optimizadas

“…Dentro de los métodos de solución de trayectoria aplicados al PPSP es posible encontrar metaheurísticas basadas en GRASP, buscando minimizar el riesgo acumulado del portafolio para un horizonte de tiempo fijo (Mira et al, 2012); Metaheurísticas de Randomized Prioriry Search (RPS) para la selección de portafolio de proyectos con restricciones de recursos dinámicas con ventanas de tiempo (Garcia, 2014), algoritmos de Tabu Search (TS) para la selección y programación simultanea de proyectos buscando maximizar el beneficio total esperado de la ejecución del portafolio, respetando interdependencias y fechas de terminación para el portafolio a ejecutar en un horizonte dado (Kumar et al, 2019), entre otros.…”