Problem instances for robust spotter scheduling problem

We randomly generated problem instances for the robust spotter scheduling problem (RSSP) which are used in the working paper "Tadumadze, G., Boysen, and N., Emde, S. (2019): Robust spotter scheduling in trailer yards. Working Paper TU Darmstadt". For detailed instance generation scheme see Section 4.1 of the aforementioned work.

We generate RSSP instances for the following three different sized trailer terminals which differ in the number of dock doors \(|D|\) of the terminal and the parking positions\(|\Pi|\) in the parking lot:

• S: \(|D| = 20\); \(|\Pi| \in \{40,60,80,100\}\)
• M: \(|D| = 50\); \(|\Pi| \in \{100,150,200,250\}\)
• L: \(|D| = 200\); \(|\Pi| \in \{400,600,800,1000\}\)

The dataset contains of 800 RSSP instances each of which is described in a text-file (labelled as “<ID>.txt” where ID is a running index of instance (from 0001 to 0800)). In particular, each file contains information about the number of transport requests (i.e., jobs) \(n\), the spotter fleet size \(m\) , processing time of transport request \(j\) (\(j=1,…,n\)) \(p_j\) , completion time of transport request \(j\) (\(j=1,…,n\))  \(C_j\) and the deadheading time it takes a spotter to move from the target position of job \(j\) to the start position of job \(j'\) (\(j,j'=1,…,n\)) \(\delta_{jj'}\) (i.e., sequence-dependent setup times). Further, each file consists of additional information which describes the direction of each transport request (i.e., whether it describes a transport request of a semitrailer from parking lot to terminal or vice versa) and the index of corresponding truck (i.e., semitrailer).

Each text-file is structured as follows:

count_jobs =

<integer number that describs the number of transport requests (i.e., \(n\))>.

Count_spotters =

<integer number that describs the number of spotters (i.e., \(m\))>.

processing_times =

<vector with \(n\) integer numbers: the \(j\)-th element of the vector describes the processing time of for transport request \(j\) (i.e., \(p_j\))>.

completion_times =

<vector with \(n\) integer numbers: the \(j\)-th element of the vector describes the completion time of \(j\)-th transport request (i.e., \(C_j\))>.

weights =

<vector with \(n\) integer numbers: the \(j\)-th element of the vector describes the relative importance of transport request \(j\) (i.e., \(w_j\))>.

driving_times =

<matrix with \(n*n\) integers: each entry in \(j\)-th column and \(j'\)-th row describes the sequence-dependent setup times between jobs \(j\) and \(j'\) (\(\delta_{jj'}\))>.

*********************job_truck_relation*********************

to_terminal =

<vector with n boolean values: the \(j\)-th element of vector has a value “\(True\)”, if the \(j\)-th job stands for transport request of a semitrailer from parking lot to terminal; “\(False\)” -- otherwise>.

trucks_index =

<vector with n integer numbers: the \(j\)-th element of vector describes the index of the semitrailer of transport request \(j\)>.

The first 360 RSSP instances (i.e., 0001-0360) are generated for the computational performance test (Section 4.2) and the next 440 instances (i.e., 0361-0800) -- for the simulation study (Section 4.4).