Professional Documents
Culture Documents
Ee 2015 01spec - Issue Ikudayisi
Ee 2015 01spec - Issue Ikudayisi
Akinola Ikudayisi
AUTHORS
Josiah Adeyemo
Akinola Ikudayisi and Josiah Adeyemo (2015). Irrigation water optimization using
ARTICLE INFO
evolutionary algorithms. Environmental Economics, 6(1-si), 200-205
0 0 0
businessperspectives.org
Environmental Economics, Volume 6, Issue 1, 2015
200
Environmental Economics, Volume 6, Issue 1, 2015
design issues (Reddy and Kumar, 2007). The paper The advantages of adopting EAs in solving optimi-
is divided into three sections followed by list of zation problems are so numerous and includes: (1)
references. Section 1 deals with evolutionary algo- EAs are solid contender for issues with non-raised,
rithms. The various applications of evolutionary irregular and multimodal functions; (2) EAs do not
algorithms in irrigation water optimization are pro- need to consider whether a function is convex, con-
vided in section 2. Conclusion of the study is pro- cave or continuous. It solves all functions without
vided in the final section. any hitch (Sarker and Ray, 2009). (3) EAs are very
ideal for solving multi-objective optimization prob-
1. Evolutionary algorithms
lems because it can handle the many conflicting
Evolutionary algorithms (EAs) are well renowned objective functions and also bring about lots of op-
optimization tools suitable and useful for searching timal solutions in a single run (Sarker,
feasible decision space and solving diverse chal- Kamruzzaman and Newton, 2003).
lenges that relates with planning, design and man-
The most popular of the EAs are Genetic algorithms
agement of natural resources (Whitley, 2001). It em-
which is a search algorithm that works based on the
ploys the method of evolution to unravel adequate
theory of natural genetics (Azamathulla et al.,
solutions that are commensurate with the challenging
2008). Genetic algorithm technique is robust in its
and complicated resource allocation problems around
capacity to search for optimal solutions and widely
the world. EAs use the theory of Charles Darwin’s
used in the optimization of water resources benefits
natural selection to search for optima solutions in a
(Arunkumar and Jothiprakash, 2013). It was devel-
given problem and they have been adopted over the
oped in 1970 and had since been accepted as a po-
years to solve diverse application problems
werful optimization method (Azamathulla et al.,
(Adeyemo, Bux and Otieno, 2010). Another interest-
2008). Examples of great research works done on
ing feature of EAs is their ability to solve multi-
multi-reservoir water optimization using Genetic
objective optimization problems (MOOP) and this
algorithms include Anwar and Clarke (2001),
has actually popularised it in the last few decades
Wardlaw and Bhaktikul (2004), Azamathulla et al.
(Adeyemo, Bux and Otieno, 2010). EAs go for dis-
(2008), Casadesús et al. (2012), Elferchichi et al.
covery of the optima from a population of points in
(2009), Bieupoude, Azoumah and Neveu (2012),
parallel rather than from a single point. These advan-
Chang et al. (2010).
tages have promoted their suitability in handling
complex design issues (Reddy and Kumar, 2006). Apart from GA, other effective and highly used
evolutionary algorithms are differential evolution
The general procedures of EAs as outlined by Eiben
(DE) algorithm, Genetic Programming (GP), evolu-
and Smith (2003) are initialization, mutation, cross-
tion strategies (ES), particle swarm optimization
over and selection. Populations of individuals which
(PSO). DE technique was created and developed by
are potential solutions are first randomly generated.
Storn and Price (1995) but in solving multi-reservoir
Each solution is assessed by using a fitness function.
system optimization problems; genetic algorithm is
A selection process is applied during each iteration
mostly adopted more than differential evolution
process to generate a new population which will be
(Goldberg, 1989). DE was firstly developed for
better than the previous population. The selection is
single objective optimization and due to its simplici-
biased towards the solution that has a better value of
ty principle and convenience in computer program-
the fitness function. During each iteration process,
ming; it has been employed for solving various ap-
the solutions undergo mutation and crossover to
plication problems (Vasan and Raju, 2007). One of
mimic the natural evolution technique. The iteration
the most popular formats of DE is the one known as
continues until convergence is reached.
DE/rand/1/bin strategy. This format of DE mainly
Over the years in the field of operations research, contains three operators: mutation, crossover and
EAs have found maximum usage in solving both selection (Singh, 2012).
single and multi-objective optimization problems
2. Application of evolutionary algorithms
(Sarker, 2009). In solving single objective optimiza-
in irrigation water optimization
tion problems, EAs always go out to obtain the best
global minimum or maximum as the case may be Considerable research works have been done; ma-
which is determined by the nature of the problem thematical models have been developed for optimiz-
being addressed (Cheng et al., 2008). On the con- ing irrigation water management for different irriga-
trary, in multi-objective optimization problems, an tion systems and reservoir systems around the
EA searches for a set of solutions that are better and world. For example, Wardlaw and Bhaktikul (2004)
fulfil the boundary conditions to the remainder solu- employed a genetic algorithm (GA) to solve the
tions in the search space. problem of irrigation scheduling and claimed better
201
Environmental Economics, Volume 6, Issue 1, 2015
solution quality by scheduling supplies as close as Paly and Zell (2009) did a comparative analysis of
possible to the Pareto front. Several other studies five Evolutionary Algorithms namely Real Valued
demonstrated the efficiency and the strength of the Genetic Algorithm, Particle Swamp Optimization,
GA approach as an optimization tool to provide Differential Evolution and two Evolution Strategy –
good solutions for an irrigation scheduling problem based algorithms on the problem of irrigation opti-
such as Reca (2001), Azamathula (2008), Fotakis mization and their result showed that both Differen-
(2012), Belaqziz (2013), Peralta (2014). tial Evolution (DE) and Particle Swamp Optimiza-
tion (PSO) are able to optimize irrigation schedules
Raju and Kumar (2004) applied GA to irrigation
and achieve results that are extremely close to the
planning problem in order to evolve efficient crop-
theoretical optimum.
ping pattern for maximum benefits for an irrigation
project in India. This methodology was adopted to A crop planning problem was formulated as a multi-
expand net profits with the imperatives, for exam- objective optimization model by Sarker and Ray
ple, progression comparison, land and water neces- (2009) and solved using three distinctive optimiza-
sities, channel limit, store stockpiling confinements tion approaches. The methodologies considered
and trimming example contemplations. The results were; Ť – constrained method, a well-known multi-
got from the GA model were contrasted with those objective evolutionary algorithm NSGA-II and their
got from Linear Programming model and they in- proposed multi-objective constrained algorithm
ferred that GA is a powerful optimization technique (MCA). They critically assessed the execution of
for irrigation water planning and can be utilized for their proposed MCA with the other two methodolo-
more intricate frameworks including non-direct gies and they broke down the arrangements from
optimization. choice making perspective. NSGA-II failed to dis-
Reddy and Kumar (2006) developed a Multi- cover plausible solutions in 69% of the cases ex-
objective Evolutionary Algorithm (MOEA) and plained. Their proposed technique MCA did more
applied same to problems involving a multipurpose excellently than NSGAII for both occasions of the
reservoir system. The methodology was developed crop planning model.
to find a set of well distributed optimal solutions In another study carried out by Chang and Chang
along the Pareto front. They employed a population (2009), a multi-objective evolutionary algorithm
based search evolutionary algorithm named Multi- named, non-dominated sorting genetic algorithm
objective Genetic Algorithm (MOGA) to overcome (NSGA-II) was applied to examine the operations of
the challenge faced by the classical methods for both Feitsui and Shihmen reservoir systems in Tai-
Multi-objective Optimization Problems (MOOP). wan. The NSGA-II was used to minimize the short-
The MOGA methodology was applied to a reasona- age indices (SI) of the two reservoirs over a long
ble reservoir system, namely Bhadra Reservoir sys- term simulation period of 49 years. Their result
tem, in India and the results obtained using the pro- demonstrated that NSGA-II is a compelling and
posed evolutionary algorithm showed that it found a vigorous multi-objective system to recognize joint
well distributed set of Pareto optimal solutions operation methodologies that will address discrimi-
along the Pareto front and hence it shows the suita- nating future maintainability needs in future.
bility of MOGA for solving multi-objective optimi- GA approaches have been successfully used for the
zation issues. identification of optimal solutions in many hydraulic
A detailed comparison to prove the superiority of problems. Elferchichi et al. (2009) developed an
evolutionary methods over classical methods was optimization model based on real-coded genetic
done by Azamathulla et al. (2008). They conducted algorithms for optimising the operation of reservoirs
a detailed comparison between two models – a Ge- in an on-demand irrigation system. The model was
netic Algorithm (GA) and Linear Programming (LP) applied and tested on the Sinistra Ofanto irrigation
and they applied it to real-time reservoir operation scheme in Italy. The model analyzed the adequacy
meant for irrigation in Chiller reservoir system in of the difference between supply and demand taking
Pradesh, India. The state variables considered by the into account the storage capacity of the reservoirs.
real-time operation model were soil moisture status They concluded that GA is an efficient algorithm for
solving problems relating to multi-reservoirs.
and the reservoir storage. The applied irrigation
depths serve as the decision variables. In a bit to A comparative analysis was carried out by Vasan
curb water wastage, the optimum crop pattern model and Raju (2009) where they compared the applica-
will only allow productive irrigation and hence, the tion of Simulated Annealing (SA), simulated
performance of both models were analyzed. GA quenching (SQ) and real-coded genetic algorithm
model gives better yield than the LP model. (RGA) to a case study of Mahi Bajaj Sugar project
202
Environmental Economics, Volume 6, Issue 1, 2015
in India. The study objective was to boost the yearly SOA). It was applied to solve a complex, non-linear,
net profits subjected to different irrigation system combined land use and water allocation problem.
constraints for 75% trustworthy stream situation. The objectives of the problem solved includes; (a)
Sensitivity investigation on different parameters The minimization of soil and groundwater pollution
utilized within the above systems showed that they and (b) the maximization of economic profit. The
yielded same solutions when compared to a set of studied area was divided into land blocks and it
ideal set of parameters. It was accordingly con- included a number of wells in fixed positions. The
cluded that SA, SQ and RGA can be used for pro- results obtained by MOSOA were compared to a
ductive solution of any irrigation system framework standard multi-objective genetic algorithm (NSGA-
with suitable constraints. II) and the former yielded better and satisfactory
outcomes as it generates a set of optimal solutions
A new evolutionary optimization method was devel- along the Pareto front and it also satisfies the com-
oped by Chen and Chang (2009) called evolutionary paction criteria.
artificial neural networks (EANN) for time series
forecasting. This optimization technique combined Another application of evolutionary algorithms is
both genetic algorithm (GA) and artificial neural demonstrated in the study carried out by Belaqziz et
network (ANN) to solve optimization problems. They al. (2013) where they proposed a new methodology
explained that the limitation of ANN is that it has for irrigation scheduling optimization based on the
inability to process a great number of information and stochastic search algorithm called Covariance Ma-
deal with non-linearity but GA has the capacity to trix Adaptation Evolution Strategy (CMA-ES). It is
supplement the inadequacy of ANN. Ordinarily, it is one of the most powerful techniques for the optimi-
very difficult to optimize the best network architec- zation of single-objective problems. It is an iterative
ture using ANN, especially for highly non-linear data stochastic optimization algorithm where at each
but the best way is to introduce other algorithms of iteration process, a population of candidate solutions
global optimization (e.g. GA) that would enhance the are sampled. They applied CMA-ES to an irrigated
search for near-optimal solution. The fundamental sector located at Tensift plain in Morocco. Their
reason for their study was to propose EANN for natu- objective was to offer the irrigation managers a
rally developing the ideal system construction model- complete scheduling tool for irrigation rounds in-
ing and association weights of ANN to the examined cluding data dates and times of opening and closing
time series. They initially investigated the execution the canals to irrigate plots and the amount of water
of the proposed EANN for the Mackey-Glass riotous needed. They concluded that the proposed approach
time arrangement and the result demonstrated that is very promising for managing and optimizing irri-
EANN has effectiveness, adequacy and its robust. gation schedules in the gravity irrigation systems.
In a study carried out by Adeyemo and Otieno Peralta, Forghani and Fayad (2014) applied Multi-
(2010), the ability of multi-objective differential objective Genetic Algorithm (MOGA) to a hydrauli-
evolutionary algorithm (MDEA) as an evolutionary cally and economically nonlinear system in which all
algorithm for solving multi- objective optimization significant flows, including stream – aquifer – reser-
problems is demonstrated. They presented an evolu- voir – diversion – return flow interactions, are simu-
tionary algorithm methodology for solving a multi- lated and optimized simultaneously for multiple pe-
objective crop planning problem. The objectives of riods. Three considered conflicting objectives are:
the problem include; minimization of total irrigation maximizing water provided from surface and
water, maximization of both the total net income groundwater resources, maximizing hydropower
from farming and the total agricultural output. They production and minimizing operation costs of moving
applied the proposed MDEA to Vaalharts irrigation water from resources to destinations. The MOGA
scheme (VIS) in South Africa and from the study, it optimizer satisfactorily generated diverse and well
was concluded that MDEA is a good algorithm for distributed solutions to show decision makers a true
solving crop planning problems. It is also an effec- picture of trade-offs between conflicting objectives.
tive and concise technique for solving multi-
objective problems in water resources systems. Conclusion
A new and innovative evolutionary algorithm de- From this study, it can be observed that many re-
veloped specifically for solving spatial optimization searchers around the globe have developed, initiated
problems was developed by Fotakis and and applied various evolutionary algorithms to solve
Sidiropoulos (2012) and it is used for solving both irrigation water problems. Also, the ability of evolu-
land use planning and resource allocation problems. tionary algorithms to evaluate multi-objective opti-
The optimization methodology is multi-objective, mization problems and find near Pareto optimal
based on non-domination criteria and it is called solutions was also demonstrated in this paper. Up
multi-objective self-organizing algorithm (MO- till now, there are few (if any) alternatives to EA-
203
Environmental Economics, Volume 6, Issue 1, 2015
based multi-objective optimization. The numerous i Which EA implementations are suited to which
applications and the rapidly growing interest in the sort of problem?
area of multi-objective evolutionary algorithms i What are the specific advantages and draw-
(MOEAs) take this fact into account. Furthermore, backs, respectively, of different techniques?
genetic algorithm (GA), as an evolutionary algo- i Are there sufficient and commonly accepted
rithm, has been utilized to solve complex non-linear definitions of quantitative performance metrics
and non-arched optimization issues. It is appropriate for multiobjective optimizers?
in achieving worldwide ideal solutions of different i The various MOEAs incorporate different con-
optimization problems as well as yielding much cepts, e.g. elitism and niching, which are in
better results when contrasted with other evolutio- principle independent of the fitness assignment
nary algorithms. The primary preference of utilizing method used. What are the benefits of these
GA is that it is chiefly suitable for remotely con- concepts? Can elitism improve multi-objective
necting the numerical display inside the optimiza- search in general?
tion model. One of the few research gaps observed
from these review is that there are no enough studies Therefore, from the review above, there are research
providing performance comparisons and investiga- gaps that have been identified as regards the opera-
tion of different aspects of the several evolutionary tion of evolutionary algorithms in solving most irri-
approaches. The few comparative studies that have gation water problems which is a multi-objective
been published remain mostly qualitative and are optimization problem. The above discussion is ad-
often restricted to a few algorithms. There is a need vantageous in providing a focus for possible appli-
to answer the following questions and observations cations of evolutionary algorithms in water re-
as regards the scope of this paper: sources management around the world.
References
1. Adeyemo, J. and Otieno, F. (2010). Differential evolution algorithm for solving multi-objective crop planning
model, Agricultural Water Management, 97 (6), pp. 848-856.
2. Adeyemo, J.A., Bux, J. and Otieno, F. (2010). Differential evolution algorithm for crop planning: Single and
multi-objective optimization model, International Journal of the Physical Sciences, 5 (10), pp. 1592-1599.
3. Anwar, A.A. and Clarke, D. (2001). Irrigation scheduling using mixed-integer linear programming, Journal of
Irrigation and Drainage Engineering, 127 (2), pp. 63-69.
4. Arunkumar, R. and Jothiprakash, V. (2013). Chaotic Evolutionary Algorithms for Multi-Reservoir Optimization,
Water Resources Management, 27 (15), pp. 5207-5222.
5. Azamathulla, H.M., Wu, F.-C., Ghani, A.A., Narulkar, S.M., Zakaria, N.A. and Chang, C.K. (2008). Comparison
between genetic algorithm and linear programming approach for real time operation, Journal of Hydro-
environment Research, 2, pp. 172-181.
6. Belaqziz, S., Fazziki, A.E., Mangiarotti, S., Le Page, M., Khabba, S., Raki, S.E., Adnani, M.E. and Jarlan, L.
(2013). An Agent based Modeling for the Gravity Irrigation Management, Procedia Environmental Sciences, 19,
pp. 804-813.
7. Belaqziz, S., Mangiarotti, S., Le Page, M., Khabba, S., Er-Raki, S., Agouti, T., Drapeau, L., Kharrou, M.H.,
Adnanai, M.E. and Jarlan, L. (2014). Irrigation scheduling of a classical gravity network based on the Covariance
Matrix Adaptation – Evolutionary Strategy algorithm, Computers and Electronics in Agriculture, pp. 64-72.
8. Bieupoude, P., Azoumah, Y. and Neveu, P. (2012). Optimization of drinking water distribution networks: Computer-
based methods and constructal design, Computers, Environment and Urban Systems, 36 (5), pp. 434-444.
9. Casadesús, J., Mata, M., Marsal, J. and Girona, J. (2012). A general algorithm for automated scheduling of drip
irrigation in tree crops, Computers and Electronics in Agriculture, 83, pp. 11-20.
10. Chang, L.-C. and Chang, F.-J. (2009). Multi-objective evolutionary algorithm for operating parallel reservoir
system, Journal of Hydrology, 377 (1), pp. 12-20.
11. Chang, L.-C., Chang, F.-J., Wang, K.-W. and Dai, S.-Y. (2010). Constrained genetic algorithms for optimizing
multi-use reservoir operation, Journal of Hydrology, 390 (1), pp. 66-74.
12. Chen, Y.-H. and Chang, F.-J. (2009). Evolutionary artificial neural networks for hydrological systems forecasting,
Journal of Hydrology, 367 (1-2), pp. 125-137.
13. Cheng, C.T., Wang, W.C., Xu, D.M. and Chau, K. (2008). Optimizing Hydropower Reservoir Operation Using
Hybrid Genetic Algorithm and Chaos, Water Resources Management, 22 (7), pp. 895-909.
14. Cui, L., Ravalico, J., Kuczera, G., Dandy, G. and Maier, H. (2011). Multi-objective Optimization Methodology for
the Canberra Water Supply System. Canberra, Australia: eWater Cooperative Research Centre.
15. Deb, K. (2001). Multi-Objective Optimization using Evolutionary Algorithms. Chichester, UK: John Wiley & Sons.
16. Eiben, A.E. and Smith, J.E. (2003). Introduction to evolutionary computing. Springer.
204
Environmental Economics, Volume 6, Issue 1, 2015
17. Elferchichi, A., Gharsallah, O., Nouiri, I., Lebdi, F. and Lamaddalena, N. (2009). The genetic algorithm approach
for identifying the optimal operation of a multi-reservoirs on-demand irrigation system, Biosystems Engineering,
102 (3), pp. 334-344.
18. Fotakis, D. and Sidiropoulos, E. (2012). A new multi-objective self-organizing optimization algorithm (MOSOA)
for spatial optimization problems, Applied Mathematics and Computation, 218 (9), pp. 5168-5180.
19. Goldberg, D.E. (1989). Genetic Algorithms in search, Optimization and Machine Learning. Mass: Addison-
Wesley.
20. Huang, Y., Li, Y.P., Chen, X. and Ma, Y.G. (2012). Optimization of the irrigation water resources for agricultural
sustainability in Tarim River Basin, China, Agricultural Water Management, 107, pp. 74-85.
21. McCarthy, A.C., Hancock, N.H. and Raine, S.R. (2014). Development and simulation of sensor-based irrigation
control strategies for cotton using the VARIwise simulation framework, Computers and Electronics in
Agriculture, 101, pp. 148-162.
22. Mishra, A.K. and Singh, V.P. (2011). Drought modeling – A review, Journal of Hydrology, 403 (1-2), pp. 157-175.
23. Mishra, B.S.P. and Dehuri, S. (2011). A Critical Survey of Single- and Multi-Objective Parallel Genetic
Algorithms, The IUP Journal of Computer Sciences, 5 (1), pp. 52-87.
24. Nkondo, M.N., van Zyl, F.C., Keuris, H. and Schreiner, B. (2004). National Water Resources Strategies, Cape
Town: Department of Water Affairs.
25. Paly, M. and Zell, A. (2009). Optimal irrigation scheduling with evolutionary algorithms. In: Proceedings of
Applications of Evolutionary Computing, (EvoWorkshops 2009). Heidelberg, Germany, Springer, pp. 142-151.
26. Peralta, R.C., Forghani, A. and Fayad, H. (2014). Multiobjective genetic algorithm conjunctive use optimization
for production, cost, and energy with dynamic return flow, Journal of Hydrology, 511, pp. 776-785.
27. Raju, K.S. and Kumar, D.N. (2004). Irrigation Planning using Genetic Algorithms, Water Resources Management,
18, pp. 163-176.
28. Reca, J., Roldán, J., Alcaide, M., López, R. and Camacho, E. (2001). Optimization model for water allocation in
deficit irrigation systems: I. Description of the model, Agricultural Water Management, 48 (2), pp. 103-116.
29. Reddy, M.J. and Kumar, D.N. (2006). Optimal Reservoir Operation Using Multi-Objective Evolutionary
Algorithm, Water Resources Management, 20 (6), pp. 861-878.
30. Reddy, M.J. and Kumar, D.N. (2007). Multiobjective differential evolution with application to reservoir system
optimization, Journal of Computing in Civil Engineering, 21 (2), pp. 136-146.
31. Sarker, R., Kamruzzaman, J. and Newton, C. (2003). Evolutionary optimization (EvOpt): a brief review and
analysis, International Journal of Computer Intelligence and Application, 3, pp. 311-330.
32. Sarker, R., Ray, T. (2009). An Improved Evolutionary Algorithm for Solving Multi-Objective Crop Planning
Models, Computers and Electronics in Agriculture, 68 (2), pp. 191-199.
33. Schütze, N., Wöhling, T., De Paly, M. and Schmitz, G. (2006). Global optimization of deficit irrigation systems
using evolutionary algorithms. In: Proceedings of Denmark, Proceedings of the XVI International Conference on
Computational Systems.
34. Singh, A. (2012). An overview of the optimization modelling applications, Journal of Hydrology, pp. 466-467,
pp. 167-182.
35. Singh, A. (2014). Irrigation Planning and Management Through Optimization Modelling, Water Resources
Management, 28 (1), pp. 1-14.
36. Storn, R.N. and Price, K.V. (1995). Differential evolution – a simple and efficient adaptive scheme for global
optimization over continuous spaces, International Computer Science Institute-Publications-TR-95-012, pp. 1-15.
37. Vasan, A. and Raju, K.S. (2009). Comparative analysis of Simulated Annealing, Simulated Quenching and
Genetic Algorithms for optimal reservoir operation, Applied Soft Computing, 9 (1), pp. 274-281.
38. Wardlaw, R. and Bhaktikul, K. (2004). Comparison of genetic algorithm and linear programming approaches for
lateral canal scheduling, Journal of Irrigation and Drainage Engineering, ASCE, 130 (4), pp. 311-317.
39. Whitley, D. (2001). An overview of evolutionary algorithms: practical issues and common pitfalls, Information
and Software Technology, 43, pp. 817-831.
205