Ant Colony Optimization for Water Resources Systems Analysis – Review and Challenges

Water resources systems analysis is the science of developing and applying mathematical operations research methodologies to water resources systems problems comprised of reservoirs, rivers, watersheds, groundwater, distribution systems, and others, as standalone or integrated systems, for single or multiobjective problems, deterministic or stochastic. The scientific and practical challenge in dealing quantitatively with water resources systems analysis problems is in taking into consideration from a systems perspective, social, economic, environmental, and technical dimensions, and integrating them into a single framework for trading-off in time and in space competitive objectives. Inherently, such problems involve modelling of water quantity and quality for surface water, groundwater, water distribution systems, reservoirs, rivers, lakes, and other systems as stand alone or combined systems. Numerous models for water resources systems analysis have been proposed during the past four decades. A possible classification for those is into: (1) methods based on decomposition in which an "inner" linear/quadratic problem is solved for a fixed low-dimension decision variables set, while that set is altered at an "outer" problem using a gradient or a subgradient technique (e.g., Alperovits and Shamir, 1977; Quindry et al. 1979, 1981; Kessler and Shamir, 1989, 1991; Eiger et al., 1994; Ostfeld and Shamir, 1996), (2) methods which utilize a straightforward non-linear programming formulation (e.g., Watanatada, 1973; Shamir, 1974; Karatzas and Finder, 1996), (3) methods based on linking a simulation program with a general nonlinear optimization code (e.g., Ormsbee and Contractor, 1981; Lansey and Mays, 1989), and (4) more recently, methods which employ evolutionary techniques such as genetic algorithms (e.g., Simpson et. al, 1994; Savic and Walters, 1997; Espinoza et al., 2005; Espinoza and Minsker, 2006), Cross Entropy (e.g., Perelman and Ostfeld, 2008), or the shuffled frog leaping algorithm (e.g., Eusuff and Lansey, 2003). Among the above classification ant colony optimization (ACO) belongs to category (4) of evolutionary techniques. Although some studies have already been conducted (e.g., Maier et al., 2003, Ostfeld and Tubaltzev, 2008; Christodoulou and Ellinas, 2010) in which ant colony optimization was utilized, the employment of ACO in water resources systems studies is still in its infancy. This book chapter reviews the current literature of ACO for water resources systems analysis, and suggests future directions and challenges for using ACO for solving water resources systems problems [parts of this Chapter are based on Ostfeld and Tubaltzev (2008), with permission from the American Society of Civil Engineers (ASCE)].