Multi-Objective Evolutionary Algorithm for Land-Use Management Problem

Abstract

Due to increasing population, and human activities on land to meet various demands, land uses are being continuously changed without a clear and logical planning with any attention to their long term environmental impacts. Thus affecting the natural balance of the environment, in the form of global warming, soil degradation, loss of biodiversity, air and water pollution, and so on. Hence, it has become urgent need to manage land uses scientifically to safeguard the environment from being further destroyed. Owing to the difficulty of deploying field experiments for direct assessment, mechanistic models are needed to be developed for improving the understanding of the overall impact from various land uses. However, very little work has been done so far in this area. Hence, NSGA-II-LUM, a spatial-GIS based multi-objective evolutionary algorithm, has been developed for three objective functions: maximization of economic return, maximization of carbon sequestration and minimization of soil erosion, where the latter two are burning topics to today's researchers as the remedies to global warming and soil degradation. The success of NSGA-II-LUM has been presented through its application to a Mediterranean landscape from Southern Portugal.

An improved knowledge-informed NSGA-II for multi-objective land allocation (MOLA)

“…Multi-objective land allocation (MOLA) can be regarded as a spatial optimization problem that aims to allocate appropriate use to certain land units subjecting to multiple objectives and constraints (Eastman, Jiang, and Toledano 1998;Datta et al 2007). MOLA shares the common difficulties and complexities of multi-objective optimization problems: the multiple non-commensurable and conflicting objectives have to be optimized simultaneously.…”

“…As a spatial optimization problem, MOLA has extra difficulties and complexities, such as the remarkably large data size, the nonlinear objective functions, and the interdependency between spatial variables. Classical optimization methods, such as enumeration, linear programing, branch and bound in integer programing, are found not feasible or appropriate to solve the MOLA problem (Datta et al 2007;Tong and Murray 2012). Meanwhile, these classical optimization methods cannot solve multi-objective optimization problems following the Pareto-based method; the multiple objectives have to be combined into a single objective function.…”

“…Heuristic algorithms, such as simulated annealing (SA), genetic/evolutionary algorithms (GAs/EAs), and particle swarm optimization (PSO) have been developed to produce approximate solutions to solve optimization problems. They have been found applicable to the MOLA problem (Aerts and Heuvelink 2002;Datta et al 2007;Duh and Brown 2007;Santé-Riveira, Crecente-Maseda, and Miranda-Barrós 2008;Cao et al 2011Cao et al , 2012Masoomi, Mesgari, and Hamrah 2013;Liu et al 2016;Sahebgharani 2016). However, since heuristic algorithms cannot guarantee optimal solutions, they suffer the risk of being trapped in local optima.…”

“…Since MOLA itself is a knowledge-informed procedure/activity, incorporating knowledge-informed rules into the heuristic optimization algorithms has been regarded as a natural and promising approach to achieve quality solutions for MOLA within a reasonable computation time (Datta et al 2007;Duh and Brown 2007;Cao et al 2012;Liu et al 2016). The knowledge-informed NSGA-II try to avoid redundant and unproductive searches by reducing the size of the search space or changing its structure.…”

“…The knowledge-informed rules were applied to decrease landscape fragmentation. Datta et al (2007) developed a series of knowledge-informed operators for the NSGA-II to solve a three-objective optimization problem in land management, including maximizing economic return, maximizing carbon sequestration, and minimizing soil erosion. The knowledge-informed rules were mainly used for encouraging compact land allocation and steering infeasible solutions to feasible ones.…”