5. Description of the simulation tool
5.1 Input and output files
5.1.1 Input files description
Discrete event simulation tool for evaluation of supply vessel schedules’ robustness and their a posteriori improvements is described in this chapter. The chapter consists of sections dedicated to description of input and output files, routing and weather modelling algorithms, interface, robustness parameters and the way they get estimated and finally multicriteria ranking algorithms.
5.1 Input and output files
In this section input and output files for the simulation tool are addressed.
5.1.1 Input files description
Input parameters
Firstly, let us address input data and its formats. Several important files are used for setting parameters of simulation. These files are: Vessels.csv, input.csv, installations_data.csv, WeatherData\\WeatherData.dat, WeatherData\\common.dat, WeatherData\\models.dat, WeatherData\\modelled_weather.dat. Their detailed description is presented below.
Vessels.csv, which contains such parameters of vessels as designed speed, maximal and minimal speeds (in knots), deadweight (in tons), all sorts of fuel consumptions and the corresponding to them costs.
Table 9. Input file format for Vessels.csv
Input.csv, which contains information about schedules of vessels including such parameters as expected arrival, discharge and departure times, and vessels that participate in
# 2
#Vessel Id Dead WeightCapacity Speed MinSpeedMaxSpeedFCCosts(kr/tonn)FCSailing(tonn/h)FCBase(tonn/h)FCInstallation(tonn/h)Start
TBN1 0 4847 1000 12 6 20 5000 0.43 0.08 0.26 16
TBN3 1 4847 1000 12 6 20 5000 0.43 0.08 0.26 16
TBN3 2 4847 1000 12 6 20 5000 0.43 0.08 0.26 16
Table 10. Input file format for Input.csv
Installations_data.csv, which contains geographical coordinates of the installations, their IDs and service durations in conditions of perfect weather, their working hours and cranes available.
Table 11. Input file format for Installations_data.csv
# N
1 19 19
Inst: Arr: Disch: Depart: Vessel: VesselID:
FBS 1.137107 1.333327 1.666667 TBN1 0
OFP 1.137107 1.333327 1.666667 TBN1 0
WEP 2.10108 2.10107 2.23858 TBN1 0
SLE 2.23927 2.29167 2.49583 TBN1 0
DRA 2.56755 2.56755 2.6613 TBN1 0
OFP 3.13711 3.33333 3.66667 TBN1 0
FBS 3.13711 3.33333 3.66667 TBN1 0
OFP 3.13711 3.33333 3.66667 TBN1 0
WEP 4.10108 4.10107 4.23858 TBN1 0
SLE 4.23928 4.29167 4.49583 TBN1 0
VOL 4.51325 4.51325 4.65909 TBN1 0
OFP 5.17427 5.33333 5.66667 TBN1 0
FBS 5.17427 5.33333 5.66667 TBN1 0
OFP 5.17427 5.33333 5.66667 TBN1 0
WEP 6.10108 6.10108 6.23858 TBN1 0
SLE 6.23928 6.29167 6.49583 TBN1 0
GLI 6.57026 6.57026 6.73693 TBN1 0
OFP 7.26368 7.33333 8.66667 TBN1 0
FBS 7.26368 7.33333 8.66667 TBN1 0
# Size
12 7
Node Id LatDeg LonDeg LayTime Open Close Cranes
FBS 0 59 5.66025 8 8 16 3
DRA 1 58.18833 2.475 2.25 7 19 1
GDR 2 58.83147 1.724972 2.9 0 24 1
GLI 3 58.7 1.666667 4 7 19 1
GRA 4 59.16433 2.485167 3 0 24 1
HDA 5 59.57333 2.228333 4.5 7 19 1
OFP 6 59.05108 5.244122 0 0 24 0
OVA 7 58.569 1.702861 2.9 0 24 1
SLE 8 58.36667 1.911667 4.9 7 19 1
TRL 9 58.63014 1.737639 2.9 0 24 1
VOL 10 58.45 1.9 3.5 0 24 1
WEP 11 58.36944 1.911111 3.3 0 24 1
sim_params.csv, which contains parameters of weather simulation for the whole time horizon to be addressed. In this file cell (2,1) sets ID of the number of 3-hour observation, at which the simulation begins (e.g. simulation begins at 12th hour of the year, then value for (2,1) should be 4), cell (2,1) sets the corresponding value for the end of the simulation, cell (2,3) sets the number of ARIMA-based weather parameters to be considered, cell (2,4) sets numbers of clusters for these parameters delimited with “;” sign, cell (2,5) sets historical horizon of the data to be considered, which should be greater or equal to the maximal order of AR or MA components of weather parameters. Cell (2,6) contains total number of replications to be carried out (note that this is an approximate number and it can be adjusted so that each sub period has the same amount of replications to be run, in case the number input already satisfies this conditions, it is exact). Cell (2, 7) contains the value to describe cluster crossing algorithm (see in the corresponding chapter below) and finally cell (2, 8) contains t for such an algorithm, cell (2, 9) contains an id of an a posteriori improvement type.
Table 12. Input file format for sim_params.csv
weatherData\\common.dat, which contains parameters of weather simulation for a current replication. In this file cell (2,1) sets ID of the number of 3-hour observation, at which the simulation begins (e.g. simulation begins at 12th hour of the year, then value for (2,1) should be 4), cell (2,1) sets the corresponding value for the end of the simulation, cell (2,3) sets the number of ARIMA-based weather parameters to be considered, cell (2,4) sets numbers of clusters for these parameters delimited with “;” sign, cell (2,5) sets historical horizon of the data to be considered, which should be greater or equal to the maximal order of AR or MA components of weather parameters.
Table 13. Input file format for common.dat
weatherData\\models.dat contains weather ARIMA models. This file is less intuitively clear, thence a more detailed description of it will be provided. So, weather parameters
#Start_Hour Finish_Hour Weather_Parameters_Count Clusters Horizon ReplicationsNumClusterCrossing Dt Impov.
2209 6576 3 3;2;2 15 100 1 0.1 0
Start_Hour Finish_Hour Weather_Parameters_Count Clusters Horizon
2192 2331 3 3;2;2 15
corresponding to i-th row weather parameter, cell (i,2) sets the cluster, where this parameter is relevant, cell(i,3) sets the ID for the beginning of time interval, during which the corresponding parameter is relevant, cell(i,4) sets the end of the corresponding period, cell (i,5) sets type of data transformation, so that “2” corresponds to the first difference of the natural logarithm of real data, “0” corresponds to the first difference of real data and “1”
corresponds to the second difference of real data, cell (i, 6) sets coefficients of AR components delimited with “;” sign, whilst their number sets the order of auto regression, cell (i,7) sets coefficients of MA component of the model identically to the way AR components are set, cell(i,8) sets mathematical expectation and standard deviation of the residuals of the corresponding model delimited with “;” sign, and finally cell (i,9) sets value of the intercept of the corresponding model.
Table 14. Input file format for models.dat
weatherData\\WeatherData.dat, which contains input data of the statistical observations of weather parameters preceding the simulation in order to have basis of both autoregressive and moving average data so as to begin the simulation.
Table 15. Input file format for WeatherData.dat
5 400 3 3;2;2 36 1000
#Weather_Parameter_TypeCluster_IDRel_Start_HourRel_Fin_HourTransf_TypeAR MA RES Intercept 0 0 0 471 2 0.737338;0.756223;-0.786544;0.1951280.546952;-0.962132;-0.5106560.000306;0.008946 0
0 0 472 982 2 0.348709;0.920283;-0.4274400.842745;-0.810342;-0.835468;-0.177236-0.00018;0.008134-0.00114
0 0 983 1998 2 0.765681;0.660037;-0.821898;0.236761;0;0;0;0;0;0;-0.0447330.071897;0.043962;0.0589460.00000488;0.0090460 0 0 1999 2927 2 0.737338;0.756223;-0.786544;0.1951280.546952;-0.962132;-0.5106560.000306;0.008946 0
0 1 983 1998 2 0.244788;0.125906;0.102729;0.676481;-0.5330690.824285;0.201968;-0.200210;-0.985554;-0.4264830.00000704;0.0093190
0 1 1999 2927 2 1.159743;-0.665595;0.145396;0;-0.099534;0;0;0.061534;-0.234008;0.1448970;0;0;0;0;0;0;0;0;0;-0.111762;0;-0.121588;-0.068267;0;0;0;0;-0.092215;0;0;0;0;0;0;0;-0.063096;0;-0.0878780.000695;0.009923 0 0 1 0 477 2 1.159743;-0.665595;0.145396;0;-0.099534;0;0;0.061534;-0.234008;0.1448970;0;0;0;0;0;0;0;0;0;-0.111762;0;-0.121588;-0.068267;0;0;0;0;-0.092215;0;0;0;0;0;0;0;-0.063096;0;-0.0878780.000695;0.009923 0 0 1 478 982 2 0.705389;0.721023;-0.742142;0.2340120.352350;-0.993921;-0.351979-0.000184;0.008280-0.00114
0 2 983 1998 2 0.449812;0;-0.196421;0;-0.049929;-0.205934;0.091838;0.793789;-0.4981710.268838;0;0.132680;0;0;0.124970;0;-0.857644;-0.1448010.000184;0.014820 0 0 2 1999 2495 2 1.943454;-1.302253;0.325664-0.979026;00.000795;0.014003 0
0 2 2496 2927 2 0.863824;0.238018;-0.2218950;-0.670524;-0.274085-0.000163;0.0137760 0 2 0 471 2 0.863824;0.238018;-0.2218950;-0.670524;-0.274085-0.000163;0.0137760
0 2 472 982 2 0.422062;0;-0.134349;-0.1196910.394490;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0.095817;0.150722-0.000578;0.0139890
1 0 0 2927 0 0.069804 -0.9986 -0.000302;0.0955790
1 1 0 2927 0 0.04916 -0.99539 -0.0000646;0.0271440
2 0 0 2927 0 0 -0.99851 -0.000252;0.0969810
2 1 0 2927 0 0.076195 -0.9984 0.000432;0.026770 0
Hour WheightCluster1WH_RES_1WheightCluster2WH_RES_2WheightCluster3WH_RES_3x_1 WDX_RES_1x_2 WDX_RES_2y_1 WDY_RES_1y_2 WDY_RES_2
0 3.227041 0 3.330134 0 2.419643 0 -0.10491 0 0.008881 0 0.061193 0 0.033882 0
3 3.321173 0 3.429464 0 2.489286 0 0.111613 0 -0.02115 0 -0.08741 0 0.027127 0
6 3.447194 0 3.490179 0 2.601786 0 -0.08107 0 0.061545 0 0.25084 0 -0.05096 0
9 3.492857 0 3.495759 0 2.621429 0 -0.01208 0 0.064934 0 -0.00948 0 -0.00063 0
12 3.444643 0 3.451786 0 2.633929 0 -0.01452 0 -0.01051 0 -0.03044 0 0.036716 0
15 3.365306 0 3.392634 0 2.591071 0 0.041869 0 0.053415 0 -0.02441 0 -0.0356 0
18 3.351786 0 3.385938 0 2.528571 0 -0.02695 0 -0.00701 0 0.111961 0 0.01404 0
21 3.295918 0 3.367857 0 2.458929 0 -0.13282 0 0.011631 0 0.043378 0 0.009097 0
weatherData\\Modeled_Weather.dat, which contains simulated weather for a current replication basing on the given set of models based in turn on a given set of input data with respect to the algorithm, presented in the especially dedicated to it section below.
Table 16. Input file format for Modeled_Weather.dat
These data files are imported into the model (into corresponding variables or expressions) before each run of the simulation by means of VBA code.
Vessels.csv, input.csv, installations_data.csv, WeatherData\\WeatherData.dat, WeatherData\\common.dat are imported at the beginning of the simulation, whereas WeatherData\\modelled_weather.dat is generated and then imported before the beginning of each replication. A sample of code for importing vessels.csv is presented in listing A-2 below.
5.1.2 Output files description