Optimization Of Regional Timber Logistics Network With Carbon Emissions Under Consideration

With the increasing awareness of environmental protection,the low carbon economy is imperative and decision makers have to make a balance between economy and emission when planning logistics network.As one of the three raw materials,only wood is green and renewable.Consequently,wood is playing a more and more important role in the construction of low carbon economy,and excellent timber logistics network and improved logistics service level are fundamental to efficient circulation of wood timber.However,general logistics network planning models are incapacity for timber logistics network planning because of its characteristics decided by features of commercial timber in circulation.In addition,differences among resolutions of optimization models at different levels and how they act in decision making are unknown yet.Therefore,it becomes the key problem faced by timber logistics industry that how to implement scientific and rational planning of timber logistics network to adapt to the trend of low-carbon economy.This study has an important theoretical value and practical value.This paper focus on optimization of regional timber logistics network with fully consideration about the characteristics of timber supply logistics.Firstly,how to measure the carbon emissions of a timber logistics network is discussed and an estimation of a 110.6 thousand cubic meters timber logistics network in Jiangle county in Fujian province in 2016 is done base on its present situation.The results are as follows:the average haul distance of timber is 45km;the total carbon emission is 783.64tCO2,where carbon emission of transportation is 694.72tC02,accounting for 88.7%and carbon emission of timber preliminary processing is 88.92tC02,accounting for 11.3%,and the unit carbon emission in the timber logistics network is 6.27kgCO2/m3?Secondly,three multi-objective optimization models at different levels from strategy to tactics are built,with two objectives are considered:cost minimization and carbon emissions minimization.All the three models developed embody the trade-off between total cost consisted of transportation cost,location cost and handling cost and total emissions caused by transportation,preliminary processing and location activities within timber logistics network,while the Pareto non-dominated solution sets express the relationships of the two objectives.According different characteristics of models,different algorithms are proposed and applied.Numerical experiments are carried out to verify validity of the models and effectiveness of the algorithm.(1)At the macro strategic level,a static multi-objective optimization mathematic model is developed,which incorporates two transportation modes:direct and transit.This model assist in logistics centers location and timber flow allocation.The normalized normal constraint method is applied to the case in Jiangle and the Pareto optimal frontier is derived.There are 4 kinds of different network structures corresponding to the curve which can support decision makers with different preference.In addition,influence on the Pareto frontier caused by different measurements of carbon emissions is investigated and it is showed that although places of Pareto optimal frontier are different,the trade-off relationship between the two objectives is almost unaffected.Then Sensitivity analyses of the Pareto optimal frontier to timber logistics network factors such as road conditions,truck types,supply level,processing loss rate and in timber logistics network planning are carried out.On average,carbon emission decreases 0.52%?1.2%while road rolling resistances decrease 1%.Objectives values decrease 2.3%while processing loss rates increase 20%.As the growing of timber supply-demand ratio increases,carbon emission decreases to a certain value,shapely at first,then slowly.Diversifying truck type combination can help reducing total cost as well as carbon emission.(2)Considering multi periodicity of timber logistics system and uncertainty demand of plants in different periods,a logistics network dynamic optimization model with multi-periods is formulated to determine the location of logistics centers and timber flow arrangement in every periods.Fuzzy theory is applied to deal with the uncertainty variables and the model is transformed into its equivalent form based on credibility theory.NSGA-II is designed and applied,in which a spanning tree-based determinant encoding is used to represent a timber transportation schedule.A crowding distance based differentiation mutation probability is adopted to derive a well-distributed non-dominated solution set.The results of numerical experiments prove that the developed model and algorithm can resolve the dynamic timber logistics network optimization problem with multi-periods in an efficient manner.Sensitivity analyses of dynamic change on supply and demand in different periods are implemented.The lowest approximate Pareto optimal frontier corresponds to scenario of unchanged,followed by increasing and decreasing,and the highest one corresponds to disordered.(3)On the view of combining strategy and tactics,incorporating timber truck routing problem into the network optimization,a timber logistics network location-routing multi-objectives optimization model is developed to determine the locations of timber logistics centers,allocation of timber flow,and timber trucks routes simultaneously.A hybrid heuristic algorithm(named NSGA-II-TS)that embeds Tabu search into NSGA-II is developed.On account of the high complexity of the problem,a special real-valued coding scheme and a corresponding modified partially mapped crossover are developed.Moreover,the third objective that measures infeasibility of chromosome is introduced.In numerical experiments,the validity of the proposed model and the effectiveness of the proposed algorithm are proved firstly.Then,application into real practice case and sensitivity analyses are conducted and results comparison with the first model is done also.Generally,the numbers of logistics structures contained in non-dominated solution sets of the location-routing model are less than the model without consideration of truck routing problem because of the trade-off effect between cost and carbon emissions in truck routing.The experimental results on practical case can unveil the trade-off between cost and emissions in timber logistics network,influences imposed on the trade-off by main influence factors and structure change of the timber logistics network caused by the compromise between the two objectives.Comparisons on the results of different models can stat the influence on the solutions of timber logistics network optimization imposed by the factors more distinctly.The findings in this research can provide some scientific guidelines and valuable references for the relevant decision makers in regional timber logistics network planning and designing.