Abstract—This paper shows the application of Ant Colony Optimization for the dispatch of crew teams for service assistance in an electricity utility. Four variations of the optimization algorithm are tested: sequential, deterministic-concurrent, random-concurrent and simultaneous. The methodology takes as input a set of service calls and the importance of assisting each of them. The algorithm is able to create the routes to be taken by each team and the sequence in which the services will be assisted taking into account the benefits of assisting a certain service as well as the impact of not assisting it. A computer program was developed to apply these methods and the results were considered better than the ones from the current methods used by the Company. Also, it is more suitable for real time daily applications. Finally, the four variations are compared. Keywords—Ant colony optimization, crew dispatch, service assistance. I. INTRODUCTION HE study presented on this paper is part of a Research and Development (R&D) Project of ANEEL (Brazilian Electricity Regulatory Agency), developed jointly by Electricity Company of Maranhão (CEMAR) and Daimon Engineering & Systems. CEMAR is a private-owned electric distribution utility, located in the northern region of Brazil which supplies over 2,000,000 customers, in the state of Maranhão, in Brazil. Currently, at CEMAR (and at many other Brazilian electric companies), just a few variables are taken into account by the Company’s dispatchers when service assistance is needed. They make their decisions based on previous knowledge, usually intuitively or by ad hoc methods. Also, the decisions are not reevaluated even if the circumstances are different. In order to select the services to be assisted, its sequence and the route to be taken by each available team, a Firstname SurnameAuthor 1 is with the National Institute of Standards and Technology, Boulder, CO 80305 USA (corresponding author’s phone: ; e-mail: ). Firstname SurnameAuthor 2 , was with Rice University, Houston, TX 77005 USA. He is now with the Department of Physics, Colorado State University, Fort Collins, CO 80523 USA (e-mail: ). Firstname SurnameAuthor 3 is with the Electrical Engineering Department, University of Colorado, Boulder, CO 80309 USA, on leave from the National Research Institute for Metals, Tsukuba, Japan (e-mail: ). consolidated meta-heuristic method was implemented: The Ant Colony Optimization (ACO) algorithm. Four variations of it were implemented and compared: sequential, deterministic- concurrent, random-concurrent and simultaneous [1]. The product developed offers to CEMAR and also to the technical community and to the society some important solutions and tools that are not yet contemplated by the current systems, making it an original project. It is a sophisticated tool that makes the most suitable decision for crew dispatch. This paper is organized in a way that the next section presents the theoretic bases of the project, the ACO meta- heuristic method. On Section III, the methodology is described. Section IV shows and discusses the results and, finally, Section V concludes the paper. II. ANT COLONY OPTIMIZATION The Ant Colony Optimization (ACO) belongs to a meta- heuristic group based on populations. This method can be used to solve the crew dispatch problem, in which there is a set of places to be visited and, in each of them, there is a prize to be taken by the visiting team. Once a team arrived at the point and receives the prize, no other team can receive it. The goal is to maximize the total prize [1]. This technique was inspired by the fact that ants from a colony guide themselves by a track of pheromones, searching for the best path to their food source. Good tracks are chosen more often, making its pheromone concentration greater as well as the likelihood of it being chosen again. However, some ants can explore other possibilities trying to find paths that are even better [1]. The problem can be presented as a graph. The service locations are the vertices and the paths are the edges. In ACO algorithms, an ant represents a solution. When constructing a solution, each ant is put on a starting point and then wanders randomly from vertex to vertex in the graph. At each vertex, an ant probabilistically selects the next vertex according to a decision policy or transition rule, which depends on the pheromone trails and on the heuristic information on the edges and vertices. Also, they deposit pheromone in the edges in order to attract other ants towards the corresponding area of the search space. The pheromones can evaporate, allowing some past history to be forgotten, and helping diversify the Comparison of Four Different Methods of the Ant Colony Optimization Algorithm Applied to Crew Dispatch for Network Services in an Electricity Utility Paulo Baumann1, Tiago Miranda1, Fabio Romero1, João Castilho Neto1, Alden Antunes1, Dário Takahata1, Leonardo Ferreira Neto1, Angelo Alves2, Luisa Azevedo2, Sérgio Valinho2 T Int'l Journal of Computing, Communications & Instrumentation Engg. (IJCCIE) Vol. 3, Issue 2 (2016) ISSN 2349-1469 EISSN 2349-1477 http://dx.doi.org/10.15242/IJCCIE.E0616010 252
5
Embed
Comparison of Four Different Methods of the Ant Colony ...iieng.org/images/proceedings_pdf/E0616010.pdf · Paulo Baumann1, Tiago Miranda1, Fabio Romero1, João Castilho Neto1, Alden
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Abstract—This paper shows the application of Ant Colony
Optimization for the dispatch of crew teams for service assistance in
an electricity utility. Four variations of the optimization algorithm are
tested: sequential, deterministic-concurrent, random-concurrent and
simultaneous. The methodology takes as input a set of service calls
and the importance of assisting each of them. The algorithm is able to
create the routes to be taken by each team and the sequence in which
the services will be assisted taking into account the benefits of
assisting a certain service as well as the impact of not assisting it. A
computer program was developed to apply these methods and the
results were considered better than the ones from the current methods
used by the Company. Also, it is more suitable for real time daily
applications. Finally, the four variations are compared.
Keywords—Ant colony optimization, crew dispatch, service
assistance.
I. INTRODUCTION
HE study presented on this paper is part of a Research and
Development (R&D) Project of ANEEL (Brazilian
Electricity Regulatory Agency), developed jointly by
Electricity Company of Maranhão (CEMAR) and Daimon
Engineering & Systems. CEMAR is a private-owned electric
distribution utility, located in the northern region of Brazil
which supplies over 2,000,000 customers, in the state of
Maranhão, in Brazil.
Currently, at CEMAR (and at many other Brazilian electric
companies), just a few variables are taken into account by the
Company’s dispatchers when service assistance is needed.
They make their decisions based on previous knowledge,
usually intuitively or by ad hoc methods. Also, the decisions
are not reevaluated even if the circumstances are different.
In order to select the services to be assisted, its sequence
and the route to be taken by each available team, a
Firstname SurnameAuthor1 is with the National Institute of Standards
and Technology, Boulder, CO 80305 USA (corresponding author’s phone:
; e-mail: ).
Firstname SurnameAuthor2, was with Rice University, Houston, TX 77005
USA. He is now with the Department of Physics, Colorado State University,
Fort Collins, CO 80523 USA (e-mail: ).
Firstname SurnameAuthor3 is with the Electrical Engineering Department,
University of Colorado, Boulder, CO 80309 USA, on leave from the National
Research Institute for Metals, Tsukuba, Japan (e-mail: ).
consolidated meta-heuristic method was implemented: The
Ant Colony Optimization (ACO) algorithm. Four variations of
it were implemented and compared: sequential, deterministic-
concurrent, random-concurrent and simultaneous [1].
The product developed offers to CEMAR and also to the
technical community and to the society some important
solutions and tools that are not yet contemplated by the current
systems, making it an original project. It is a sophisticated tool
that makes the most suitable decision for crew dispatch.
This paper is organized in a way that the next section
presents the theoretic bases of the project, the ACO meta-
heuristic method. On Section III, the methodology is
described. Section IV shows and discusses the results and,
finally, Section V concludes the paper.
II. ANT COLONY OPTIMIZATION
The Ant Colony Optimization (ACO) belongs to a meta-
heuristic group based on populations. This method can be used
to solve the crew dispatch problem, in which there is a set of
places to be visited and, in each of them, there is a prize to be
taken by the visiting team. Once a team arrived at the point
and receives the prize, no other team can receive it. The goal is
to maximize the total prize [1].
This technique was inspired by the fact that ants from a
colony guide themselves by a track of pheromones, searching
for the best path to their food source. Good tracks are chosen
more often, making its pheromone concentration greater as
well as the likelihood of it being chosen again. However, some
ants can explore other possibilities trying to find paths that are
even better [1].
The problem can be presented as a graph. The service
locations are the vertices and the paths are the edges. In ACO
algorithms, an ant represents a solution. When constructing a
solution, each ant is put on a starting point and then wanders
randomly from vertex to vertex in the graph. At each vertex,
an ant probabilistically selects the next vertex according to a
decision policy or transition rule, which depends on the
pheromone trails and on the heuristic information on the edges
and vertices. Also, they deposit pheromone in the edges in
order to attract other ants towards the corresponding area of
the search space. The pheromones can evaporate, allowing
some past history to be forgotten, and helping diversify the
Comparison of Four Different Methods of the
Ant Colony Optimization Algorithm Applied to
Crew Dispatch for Network Services in an
Electricity Utility Paulo Baumann1, Tiago Miranda1, Fabio Romero1, João Castilho Neto1, Alden Antunes1, Dário