IEEE CIS Task Force on Multi-modal Optimization


Population or single solution search-based optimization algorithms (i.e. meta,hyper-heuristics) in their original forms are usually designed for locating a single global solution, despite the existence of multiple optima in the search space. Representative examples include among others evolutionary and swarm intelligence algorithms. These search algorithms typically converge to a single solution because of the global selection scheme used. Nevertheless, many real-world problems are "multi-modal" by nature, i.e., multiple satisfactory solutions exist. In the presence of multiple global and local optimal solutions in a problem, it may be desirable to locate many such "satisfactory" solutions, or even all of them, so that a decision maker can choose one that is most proper in his/her problem domain. Identifying multiple optimal solutions might also provide valuable, and innovative, insights to the decision maker about the properties and the structure of the problem at hand, which is essential for a real-world optimization problem.

As such, given a multi-modal problem with multiple local/global optimal solutions, the main objective of a multi-modal optimization task is to locate as many as possible optimal solutions (global and local), in order to help the decision maker to acquire better knowledge about the different available optimal solutions of the problem at hand.

Numerous techniques have been developed in the past for locating multiple optima (global and/or local). These techniques are commonly referred to as "niching" methods. A niching method can be incorporated into a standard search-based optimization algorithm, in a sequential or concurrent way, to locate multiple optimal or suboptimal solutions. Sequential approaches locate optimal solutions progressively over time, while concurrent approaches promote and maintain formation of multiple stable subpopulations within a single population. Classic niching techniques include crowding, fitness sharing, derating, restricted tournament selection, clearing, speciation, etc. In more recent times, niching methods have also been developed for meta-heuristic algorithms such as Particle Swarm Optimization, Differential Evolution and Evolution Strategies.

Despite niching techniques first appearing more than 30 years ago (in the 1980s), we consider now niching techniques (or multi-modal optimization) is re-surging as an increasingly important research topic, attracting researchers from across a wide range of research fields, including Evolutionary Computation (EC) and Swarm Intelligence (SI). Considering its general applicability to a wide range of practical applications, multi-modal optimization can be identified and/or adopted in various optimization tasks such as continuous, combinatorial/discrete, constrained, dynamic, multi-objective, and bi-level optimization problems.


The key objective of this Task Force is to promote research on multi-modal optimization, including its development, education and understanding of sub topic areas of multi-modal optimization.

The main objectives of the task force can be summarized as follows:

  • create an active and healthy community to promote theme areas of multi-modal optimization
  • make student, researchers, end-users, developers, and consultants aware of the state-of-the-art
  • promote the use of multi-modal methodologies/techniques and tools
  • organize of conferences/workshop with IEEE CIS Technical Co-Sponsorship
  • organize tutorials, workshops and special sessions
  • launch edited volumes, books and special issues in journals

Anticipated interest

This task force will focus on all aspects of multi-modal optimization, including theory, practice and applications covering all different search-based paradigms, such as Evolutionary Computation and Swarm Intelligence algorithms.

Topics of interest include but are not limited to the following:

  • Benchmarking multi-modal optimization methods, including test problem design, analysis and performance metrics
  • Comparative studies of various multi-modal optimization methods
  • {Hyper,Meta}-heuristic approaches for multi-modal optimization problems
  • Handling the issue of niching parameters in multi-modal optimization methods (adaptive or parameter-less methods)
  • Handling the scalability (both dimensionality and modality) issue in multi-modal optimization methods
  • Hybridization of meta-heuristic approaches with neural networks, fuzzy systems, information theory, statistics, mathematical modeling, etc., for multi-modal optimization
  • Landscape analysis for multi-modal optimization problems
  • Multi-modal optimization methods that incurs lower computational costs (minimal budgets)
  • Multi-objective approaches for multi-modal optimization methodologies/problems
  • Multi-modal optimization approaches for multi-objective methodologies/problems
  • Multi-modal optimization in bi-level optimization problems.
  • Multi-modal optimization in combinatorial/discrete optimization problems
  • Multi-modal optimization in computational expensive optimization problems
  • Multi-modal optimization in constrained optimization problems
  • Multi-modal optimization in dynamic environments
  • Multi-modal optimization in large-scale optimization problems.
  • Multi-modal optimization methods applied to engineering and other real-world optimization problems
  • Multi-modal optimization methods using parallel or distributed computing techniques
  • Multi-modal optimization methods to locate all local and global optima
  • Novel level-set, niching, basin of attraction identification methodologies
  • Exploration vs exploitation in multi-modal optimization
  • Operational Research approaches for multi-modal optimization problems
  • Theoretical analysis and developments in multi-modal optimization



Past Activities



Michael G. Epitropakis (Chair)
Data Science Institute,
Department of Management Science,
Lancaster University Management School,
Lancaster University,
Lancaster, United Kingdom.

Andries Engelbrecht (Vice-Chair)
South African Research Chair in Artificial Intelligence,
Department of Computer Science,
School of Information Technology,
University of Pretoria,
Pretoria 0002, South Africa.

Xiaodong Li (Vice-Chair)
School of Computer Science and Information Technology,
RMIT University,
Melbourne, VIC 3001, Australia.

Member List:

Carlos A. Coello Coello, CINVESTAV-IPN, Mexico
Kalyanmoy Deb, Michigan State University, USA
Andries Engelbrecht, University of Pretoria, South Africa
Michael G. Epitropakis, Lancaster University, UK
Jonathan Fieldsend, University of Exeter, UK
Jian-Ping Li, Bradford University, UK
Xiaodong Li, RMIT University, Australia
Jonathan Mwaura, University of Pretoria, South Africa
Konstantinos Parsopoulos, University of Ioannina, Greece
Vassilis Plagianakos, University of Thessaly, Greece
Mike Preuss, University of Munster, Germany
Bruno Sareni, Universite de Toulouse, INP-ENSEEIHT/LAPLACE, France
Ofer M. Shir, Tel-Hai College and MIGAL Institute, Israel
Patrick Siarry, Universite Paris-Est Creteil Val-de-Marne, France
P. N. Suganthan, Nanyang Technological University, Singapore
Michael N. Vrahatis, University of Patras, Greece
Simon Wessing, TU Dortmund, Germany




Contact us

If you have any suggestions for this task force, please contact: Michael Epitropakis.