Empirical Study of Multi-objective Ant Colony Optimization
to Software Project Scheduling Problems
Jing Xiao
School of Computer Science
South China Normal University
Guangzhou 510631, China
xiaojing@scnu.edu.cn
Mei-Ling Gao
School of Computer Science
South China Normal University
Guangzhou 510631, China
galy108@163.com
Min-Mei Huang
School of Public Administration
South China Normal University
Guangzhou 510006, China
huangmm@scnu.edu.cn
ABSTRACT
The Software Project Scheduling Problem (SPSP) focuses on the
management of software engineers and tasks in a software project
so as to complete the tasks with a minimal cost and duration. It’s
becoming more and more important and challenging with the
rapid development of software industry. In this paper, we employ
a Multi-objective Evolutionary Algorithm using Decomposition
and Ant Colony (MOEA/D-ACO) to solve the SPSP. To the best
of our knowledge, it is the first application of Multi-objective Ant
Colony Optimization (MOACO) to SPSP. Two heuristics capable
of guiding the algorithm to search better in the SPSP model are
examined. Experiments are conducted on a set of 36 publicly
available instances. The results are compared with the
implementation of another multi-objective evolutionary algorithm
called NSGA-II for SPSP. MOEA/D-ACO does not outperform
NSGA-II for most of complex instances in terms of Pareto Front.
But MOEA/D-ACO can obtain solutions with much less time for
all instances in our experiments and it outperforms NSGA-II with
less duration for most of test instances. The performance may be
improved with tuning of the algorithm such as incorporating more
heuristic information or using other MOACO algorithms, which
deserve further investigation.
Categories and Subject Descriptors
I.2.8 [Artificial Intelligence]: Problem Solving, Control Method,
Search-Heuristic methods.
General Terms
Algorithms, Performance, Experimentation, Management.
Keywords
Software Project Scheduling; Multi-objective Optimization; Ant
Colony Optimization; Automatic Software Project Management;
MOEA/D; NSGA-II.
1. INTRODUCTION
Nowadays, with the booming development of software industry,
the software project is becoming more and more complicated. To
make a project successful, a software project manager must
consider how to plan, schedule, and monitor tasks and employees
more efficiently. However, it’s very difficult and time-consuming
for people to manage these tasks manually. Therefore, it’s
valuable to pay special attention to the research of computer aided
tools to improve the software project management. Generally, the
Software Project Scheduling Problem (SPSP) [1] focuses on
finding an optimal allocation of employees to tasks according to
the resources and Task Precedence Graphs (TPGs). SPSP is NP-
hard [2].
Ant colony optimization (ACO) is a bionic intelligence
optimization algorithm that was initially proposed by Dorigo [3]
for handling single-objective optimization problems. ACO has
been applied to a lot of optimization problems in operations
research such as resource constraint project scheduling problems
[4], flow shop problems [5, 6] and job shop scheduling problems
[7]. The employee allocation model of SPSP was first designed by
Alba and it has been solved by Genetic Algorithm (GA) in [1].
Later Chicano [8] solved the SPSP using multi-objective
metaheuristic algorithms including NSGA-II, SPEA2, PAES,
MOCell and GDE3. And Luna [9] performed a scalability analysis
of multi-objective metaheuristics. In addition, there are a few
papers to solve the SPSP with Ant Colony Optimization (ACO).
Xiao and Ao [2] proposed the original application of single-
objective combinatorial ant colony optimization to SPSP.
Afterwards, Crawford [10] proposed a max-min ant system
algorithm to solve SPSP.
As ACO has been successfully applied to a variety of
combinatorial optimization problems [11], it has soon been
extended to multi-objective optimization problems [12, 13].
Multi-objective optimization problems are those with several,
typically conflicting criteria for evaluating solutions. Even though
we can’t define the uniform priori preference of the objectives in
all environments, we can try to get some optimal solutions that
might be the best in a special case. Therefore, differing from
single-objective optimization problems, each multi-objective
optimization algorithm maintains a non-dominated archive and
outputs all the non-dominated solutions at the end of the algorithm.
Many people have been working on the Multi-Objective Ant
Colony Optimization (MOACO). The literature can be divided
into two categories: the one is MOACO algorithms and their
applications [12,13,14,15,16,17,18]; the other is the analysis of
alternative design choices [19,20,21] of MOACO. There are many
classical MOACO algorithms, such as BicriterionAnt [17],
Multiple Ant Colony System [14], Pareto Ant Colony
Optimization [15], COMPETants [16] and several mACO
Variants [18]. In addition, with the successful application of
Multi-Objective Evolutionary Algorithms based on
Decomposition (MOEA/D) [22], the MOEA/D-ACO [23] has
been proposed. MOEA/D-ACO is an algorithm framework that
decomposes a multi-objective optimization problem into series of
single-objective optimization problems where each ant is used to
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that copies
bear this notice and the full citation on the first page. Copyrights for
components of this work owned by others than ACM must be honored.
Abstracting with credit is permitted. To copy otherwise, or republish, to
post on servers or to redistribute to lists, requires prior specific permission
and/or a fee. Request permissions from Permissions@acm.org.
GECCO '15, July 11 - 15, 2015, Madrid, Spain
© 2015 ACM. ISBN 978-1-4503-3472-3/15/07…$15.00
DOI: http://dx.doi.org/10.1145/2739480.2754702