Resource-controlled stochastic customer order scheduling via particle swarm optimization with bound information
| DOI | https://doi.org/10.1108/IMDS-02-2022-0105 |
| Published date | 19 July 2022 |
| Date | 19 July 2022 |
| Pages | 1882-1908 |
| Subject Matter | Information & knowledge management,Information systems,Data management systems,Knowledge management,Knowledge sharing,Management science & operations,Supply chain management,Supply chain information systems,Logistics,Quality management/systems |
| Author | Yaping Zhao,Xiangtianrui Kong,Xiaoyun Xu,Endong Xu |
Resource-controlled stochastic
customer order scheduling via
particle swarm optimization with
bound information
Yaping Zhao and Xiangtianrui Kong
Department of Transportation Economics and Logistics Management,
College of Economics, Shenzhen University, Shenzhen, China
Xiaoyun Xu
Department of Operations and IT, Ateneo Graduate School of Business,
Ateneo de Manila University, Makati City, Philippines, and
Endong Xu
Department of Transportation Economics and Logistics Management,
College of Economics, Shenzhen University, Shenzhen, China
Abstract
Purpose –Cycle time reduction is important for order fulling process but often subject to resource constraints.
This study considers an unrelated parallel machine environment where orders with random demands arrive
dynamically. Processing speeds are controlled by resource allocation and subject to diminishing marginal
returns. The objective is to minimize long-run expected order cycle time via order schedule and resource
allocation decisions.
Design/methodology/approach –A stochastic optimization algorithm named CAP is proposed based on
particle swarm optimization framework. It takes advantage of derived bound information to improve local
search efficiency. Parameter impacts including demand variance, product type number, machine speed and
resource coefficient are also analyzed through theoretic studies. The algorithm is evaluated and benchmarked
with four well-known algorithms via extensive numerical experiments.
Findings –First, cycle time can be significantly improved when demand randomness is reduced via better
forecasting. Second, achieving processing balance should be of top priority when considering resource
allocation. Third, given marginal returns on resource consumption, it is advisable to allocate more resources to
resource-sensitive machines.
Originality/value –A novel PSO-based optimization algorithm is proposed to jointly optimize order schedule
and resource allocation decisions in a dynamic environment with random demands and stochastic arrivals. A
general quadratic resource consumption function is adopted to better capture diminishing marginal returns.
Keywords Stochastic customer order, Resource-controlled scheduling, Unrelated parallel machine, Particle
swarm optimization, Bound information
Paper type Research paper
1. Introduction
Improving responsiveness in customer order fulfilment process is crucial for businesses to
win competitive advantages. As one of the most important measures for responsiveness,
IMDS
122,8
1882
Funding: This work was supported by the National Science Foundation of China under Grant 72001145;
the Ministry of Education Program in Humanities and Social Sciences under Grant 20YJC630226; the
National Science Foundation of Guangdong Province under Grant 2022A1515011235; and the New
Teacher Research Start-up Foundation of Shenzhen University under Grant 00000270. The funders
support the analysis, writing and submission of the work. The authors have no competing interests, and
thank anonymous reviewers for their valuable suggestions and comments.
The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/0263-5577.htm
Received 21 February 2022
Revised 10 May 2022
Accepted 29 June 2022
Industrial Management & Data
Systems
Vol. 122 No. 8, 2022
pp. 1882-1908
© Emerald Publishing Limited
0263-5577
DOI 10.1108/IMDS-02-2022-0105
cycle time has been under the spotlight ever since the introduction of the modern
manufacturing industry and remains a very active area of research even today (Godinho
Filho and Saes, 2013). Reducing cycle time not only improves the asset turnover and cash
flow, but also frees up machines and makes it possible to respond to new unexpected
requirements. In a dynamic environment with varying demands, production managers have
to coordinate available resources in the factory and make timely decisions to ensure all
incoming customer orders are manufactured in a fast and relatively stable manner.
At the micro level, cycle time of a machine station is made up of the sum of the following
two components:
(1) Processing Time. The processing time is primarily affected by processing
environment (single machine, parallel machines, workshops, etc.) as well as the
resource allocated to the processing facility. In many industries, such as the paper
industry (Leung et al., 2005a), pharmaceutical industry (Leung et al., 2005b) and
semifinished lenses industry (Ahmadi et al., 2005), the processing environment
usually consists of multiple machines running in parallel but with heterogeneous
speeds for different product types. In addition, to maintain flexibility in handling
customer orders and enhance the coordination of various machines, the machine
processing capability can usually be controlled by allocating available resources.
These resources include but are not limited to money, energy, catalysts, overtime,
subcontracting or additional manpower to manage operations (Shabtay and Steiner,
2007;Edis et al., 2013).
(2) Queue Time. In most simple processing environment, the queue time is affected
primarily by the uncertainties in arrival pattern and demand quantity. For customer
order scheduling problems in particular, one also needs “wait-to-match”time, which
is the internal waiting time for the every component of an order to be ready in order to
make sure the customer order leaves the processing facility in a single shipping. In a
dynamic system, it is common that queue time comprises 90 percent or more of total
cycle time (Hopp and Spearman, 2008).
However, considering resource allocation in a stochastic environment is very challenging to
implement with widely applied production management information systems such as MRP
and MRP-II. The main difficulties include but are not limited to the following aspects. First,
when handling orders in a dynamic environment, managers are usually lack of accurate
demand information, such as order arrival time and demand quantity in advance (Sato et al.,
2020). Such randomness in order information contradicts the inherent assumption in MRP-
like systems on steady demand. Without accurate information on order amount and arrival
time, it is difficult for the production system to make effective planning and scheduling
decisions. Second, each incoming customer order can contain many different product types
that must be delivered as one single shipment. Such synchronization of arrival and departure
requires complex coordination among machines. Additional constraints such as load
balancing and internal queuing time reduction are usually not automatically taken into
account by the production management system. Finally, when allocating the limited
resources, decision makers must take into account of the heterogeneity in machines and the
sensitivity of their processing speeds on resources, which adds another layer of machine
intercoupling and further complicates the problem.
To tackle these difficulties, this paper considers customer order scheduling and resource
allocation in a dynamic environment with heterogeneous parallel machines. Customer orders
arrive stochastically, and the amounts of various products in each customer order are
random. Orders are processed by heterogeneous parallel machines whose processing speeds
are adjusted through resource allocation. The available resources are limited and could have
Stochastic
customer order
scheduling
1883
To continue reading
Request your trial