2.5 Optimization Objectives
2.5.4 Optimization Parameters of Mobile Environments
opti-mization objectives introduced in Sections 2.5.1, 2.5.2, and 2.5.3, this section examines the relations between the entities and how these relations can be parameterized to build different scenarios for the operation of a mobile en-vironment.
Since regions are key entities in mobile environments a brief overview of their structure and implications is given first. The regionalization influences the peculiarity of the criteria bound to the optimization objectives introduced above. Based on these insights measures for the alteration of the regional-ization are developed.
It was already stated in Definition 2.13 that a region is defined by the assets and workers associated to it and by the location of its depot. It is thus feasible
36 CHAPTER 2. DOMAIN MODEL OF MOBILITY to state that a region is a compound of assets, workers, and a depot. This coherency will be utilized to model the optimization parameters of mobile environments.
For the sake of manageability mobile organizations often split their area into several regions with independent responsibility for the business processes performed. Examples of such a setup of a German power supply are depicted in Figures 5.2 and 5.3 (p. 139) where the latter shows the detail marked in the former. As already introduced in Section 1.1.2, regions are subject to historical evolution, which lets examining them be a promising intention to improve the cost situation.
The following parameters influencing the performance of mobile environ-ments were identified with respect to the domain model. They can be used to find potential changes to the mobile environment, which improve the per-formance of process execution.
Location of the Depot (LD) Given are the regions and the assets as-signed to each region. Unknown are the locations of the depots of the regions. The location of the depot of a region r ∈ R can be constituted in different ways. Choose the location ldepotr
LD1 as the center of gravity of the assets e ∈ Er of the region r ∈ R.
LD2 such that the sum of the distances between the depot and the assets $e∈E
rdist(ldepotr , e) is minimal.
LD3 by visual selection (e.g. mouseclick).
Closest Depot (CD) Given are the locations of the depots and the assets.
Unknown is the assignment of assets to regions, i.e. the borders of the regions. Assets e ∈ E are assigned to regions such that each asset is assigned to the depot closest to it. This criterion can either be applied after relocating the depots with LD or independently. The assets are assigned to regions such that the distance dist(ldepotr , e) is minimal for all assets e ∈ E and regions r ∈ R. The regions can be found as the Voronoi tessellation (see, e.g., [85]) of the depots.
Equal Average Travel Effort (EATE) Given are the locations of the de-pots and the assets. Unknown is the assignment of assets to regions, i.e. the borders of the regions. Assets e ∈ E are assigned to regions such that the average travel effort in all regions is equal. This criterion can either be applied after relocating the depots with LD or independently.
The assets are assigned to regions such that the weighted sum
$
e∈Erdist(lrdepot, e)
|Wr| is equal for all regions r ∈ R.
2.5. OPTIMIZATION OBJECTIVES 37 Weighted Average Travel Effort (WATE) Given are the locations of the depots, the assets, and the number nvisitse ∈ Q of visits per as-set and time unit. Unknown is the assignment of asas-sets to regions.
Assets e ∈ E are assigned to regions such that the average travel effort in all regions is equal. This criterion can either be applied after relo-cating the depots with LD or independently. The assets are assigned to regions such that the sum $e∈Erdist(lrdepot, e) nvisitse is equal for all regions r ∈ R.
Equal Asset-Worker Ratio (EAWR) Given are the regions with depots and assets assigned and the workers. Unknown is the assignment of workers to the regions. Workers w ∈ W are assigned such that the quotient |W|Er|
r| is equal for all regions r ∈ R.
Equal Qualification Distribution (EQD) Let nra be the number of exe-cutions of the activity a ∈ A in the region r ∈ R in a given period of time (e.g. one year). Let further
qreqr = ("
a∈A
(nraqreqa1 ), . . . ,"
a∈A
(nraqareqm)), m = |Q|
be a |Q|-tuple representing how often each qualification is required in this region per year. Let further
qravail = ( "
w∈Wr
qavailw1 , . . . , "
w∈Wr
qavailwm ), m = |Q|
be a |Q|-tuple representing how often each qualification is available in region r ∈ R. The workers w ∈ W are assigned to the regions such that the element-wise quotients
qrreq
qravail are equal for all regions r ∈ R.
Total Number of Regions (TNR) Given are the assets. Based on the other criteria regions and depots are created from scratch. A Voronoi diagram with the assets being the Voronoi sites may be used to support the creation of the regions.
Start Location (SL) Given are the regions with assets, workers, and de-pots assigned. Given is further the home location of each worker.
Choose the location where workers start their tour SL1 to be the depot of the region r ∈ R.
38 CHAPTER 2. DOMAIN MODEL OF MOBILITY SL2 to be the home location of each worker. In this scenario a regular (e.g. weekly) meeting at the depot is usually necessary for social and business reasons and must be considered. The outcome can be compared to additional costs, e.g. additional cars and commu-nication costs.
Qualification Dependent Scheduling (QDS) In environments with dy-namic scheduling short response times might be of high importance.
If specialists with unique qualifications exist they only might get work assigned close to the center of their region or close to equipment known for frequent emergencies.
Process Execution Frequency (PEF) Processes might be executed more often than the respective legal restrictions demand. Reducing the case count might decrease overall cost. In turn the repair costs of the respective assets may rise. The criterion variates the number of executions for certain process types.
Additional Qualifications (AQ) Workers get qualifications that accel-erate their throughput or qualifications that are frequently required above availability. Outcomes can be compared to qualification costs.
Scheduler Heuristic Selection (SHS) Different scheduling heuristics are utilized for
SHS1 certain regions.
SHS2 the whole organization.
Note that an enterprise may define other or more parameters of their choice.
In Chapter 5 a selection of these parameters is used for the validation of this work.