An Agile Adaptive Biased-Randomized Discrete-Event Heuristic for the Resource-Constrained Project Scheduling Problem

Xabier A. Martin,Rosa Herrero,Angel A. Juan,Javier Panadero
DOI: https://doi.org/10.3390/math12121873
IF: 2.4
2024-06-17
Mathematics
Abstract:In industries such as aircraft or train manufacturing, large-scale manufacturing companies often manage several complex projects. Each of these projects includes multiple tasks that share a set of limited resources. Typically, these tasks are also subject to time dependencies among them. One frequent goal in these scenarios is to minimize the makespan, or total time required to complete all the tasks within the entire project. Decisions revolve around scheduling these tasks, determining the sequence in which they are processed, and allocating shared resources to optimize efficiency while respecting the time dependencies among tasks. This problem is known in the scientific literature as the Resource-Constrained Project Scheduling Problem (RCPSP). Being an NP-hard problem with time dependencies and resource constraints, several optimization algorithms have already been proposed to tackle the RCPSP. In this paper, a novel discrete-event heuristic is introduced and later extended into an agile biased-randomized algorithm complemented with an adaptive capability to tune the parameters of the algorithm. The results underscore the effectiveness of the algorithm in finding competitive solutions for this problem within short computing times.
mathematics
What problem does this paper attempt to address?
The paper attempts to address the problem of how to effectively schedule tasks in a project under resource constraints to minimize the total project completion time (i.e., minimize the project duration). Specifically, the paper focuses on the Resource-Constrained Project Scheduling Problem (RCPSP), which is an NP-hard problem involving multiple tasks with time dependencies and resource-sharing constraints. The paper proposes a new Discrete-Event Heuristic (DEH) algorithm and optimizes its performance by introducing adaptive mechanisms and biased-randomization techniques, thereby finding competitive solutions within a short computation time. ### Background of the Paper In large manufacturing industries such as aircraft or train manufacturing, companies typically need to manage multiple complex projects. Each project consists of multiple tasks that share limited resources and often have time dependencies. A common objective is to minimize the total project completion time (i.e., minimize the project duration) while satisfying the constraints of task precedence and resource availability. ### Problem Description The core issues of RCPSP are: 1. **Task Scheduling**: Determine the start time of each task so that all tasks can be completed while satisfying time dependencies and resource constraints. 2. **Resource Allocation**: Allocate limited resources reasonably to ensure that each task can obtain the required resources during execution. 3. **Minimize Project Duration**: Shorten the total project completion time as much as possible under the above constraints. ### Solution The paper proposes a new Discrete-Event Heuristic (DEH) algorithm, further extended to an Agile Biased-Randomized DEH (BR-DEH) algorithm, combining adaptive parameter adjustment mechanisms. The main contributions include: 1. **Efficiency**: Provide reasonable solutions within a short time (even for larger instances, computation time is less than 1 second). 2. **Adaptability**: Dynamically adjust algorithm parameters through a self-learning mechanism without the need for pre-analysis or incurring expensive computational costs. 3. **Local Search Improvement**: Introduce an extended Forward-Backward Improvement (FBI) method to enhance solution quality. 4. **Path Relinking**: Further improve the quality of generated solutions through a path relinking process. ### Method Overview 1. **Initial Solution Generation**: Generate an initial solution using the biased-randomized DEH algorithm. 2. **Local Search**: Improve the initial solution through the extended FBI method. 3. **Multi-Start Process**: Generate new solutions multiple times, adjusting biased-randomization parameters and sorting strategies in each iteration. 4. **Path Relinking**: Explore trajectories between high-quality solutions using the path relinking process to further enhance solution quality. ### Conclusion The paper experimentally verifies the effectiveness and efficiency of the proposed BR-DEH algorithm in solving the RCPSP problem, capable of finding competitive solutions within a short computation time.