Abstract
Among the emerged metaheuristic optimization techniques, ant colony optimization (ACO) has received considerable attentions in water resources and environmental planning and management during last decade. Different versions of ACO have proved to be flexible and powerful in solving number of spatially and temporally complex water resources problems in discrete and continuous domains with single and/or multiple objectives. Reviewing large number of peer reviewed journal papers and few valuable conference papers, we intend to touch the characteristics of ant algorithms and critically review their state-of- the-art applications in water resources and environmental management problems, both in discrete and continuous domains. The paper seeks to promote Opportunities, advantages and disadvantages of the algorithm as applied to different areas of water resources problems both in research and practice. It also intends to identify and present the major and seminal contributions of ant algorithms and their findings in organized areas of reservoir operation and surface water management, water distribution systems, urban drainage and sewer systems, groundwater managements, environmental and watershed management. Current trends and challenges in ACO algorithms are discussed and called for increased attempts to carry out convergence analysis as an active area of interest.
1 Introduction
As the spatial and temporal complexity of the water resources management and environmental problems increases, application of metaheuristic algorithms extends dramatically. Since the introduction of ACO (Dorigo et al. 1996), Different versions and refinements to the original ant algorithm are proposed and applied to solve various problems in water resources and environmental management. Ostfeld (2011) presented a review paper on ACO for water resources systems analysis. Although his attempt is acknowledged, it does not fully address the advances made in continuous and multiple objective ACO, recent wide range applications of different versions of ACO, and the enhancements in their convergences and constraint handlings.
6 Current Trends in ACO
While both the performance of ACO algorithms and our theoretical understanding of their working have significantly increased, there are several areas in which only preliminary steps have been taken and much more research are to be done. Extension of ACO algorithms to more complex optimization problems such as (1) dynamic problems, (2) stochastic problems, and (3) multiple objective problems is known to be the first area. Other active research directions in ACO include the effective parallelization of ACO algorithms and understanding and characterization of the behavior of ACO algorithms and their convergences while solving a novel problem (Dorigo and Stutzle 2004).
