ML20008D456

Applicant Exhibit A-86,consisting of Excerpt of 751219 Document Entitled, Application of Computer-Aided Dispatch in Law Enforcement-Introductory Planning Guide, Prepared for DOJ

ML20008D456
Person / Time
Site:	Seabrook
Issue date:	06/16/1989
From:	Garcia E, Gurfield R, Sohn R CALIFORNIA INSTITUTE OF TECHNOLOGY, PASADENA, CA
To:
References
OL-A-086, OL-A-86, NUDOCS 9003020258
Download: ML20008D456 (6)

Text

.

pge c;,.s &-usim -ot p f/fg gy n so.o io wa w 110 JW 18- P5 $2 DPLICJDLOF

' 00CKEilNG A SEifVlfT COMPUTER-AIDED D13 PATCH IN I.AW ENFORCEMENT

-an introductory planning guide R.L.Sohn R. M. Gurfield E. A. Garcia

'J. E. Fielding

(%

Jet Propulsion Laboratory-California Institute of Technology Pasadena, California 91103 December 19.1975 in'288M888ssha PDR

• Prepare <1 f oe

- National Criminal Justice Inf ormation and Statistics Service

(,- UNITED ST ATES DEP ARTMENT OF JUSTICE e MA4 E6

' ddhional studies of keycommand and contsol functions A

, mall segments of 5 or 10 minutes and counting (particularly dispatch stations) should be performed in order .n messages in each such segment.

?- . to provide a better underst'anding of the impact of technological i

' nnsvations on these operations. Stressful environments should e Percent utilization of dispatcher's time, For a non.

in no way be aggravated; personnel who lack good typing skills, CAD system, this will be the total of all message such as field officers temporarily assigned to dispatch duties, times plus some allowance that will have to be may find- it difficult to perform in a " type or perish" determined by observation for time spent in

' environment.~

writing, handling slips or forms, etc., without simultaneous talking or listening. '

The two departments observed and reported here are representative and provide some basis for estimating work loads e Average number of messages per case, at dispatcher. consoles. The planner should consider making similar measurements during a typical busy period, which is e Number ofincidents being handled at one time for the load that determines the major requirements of a system.

each 5 or 10 minute segment (this gives the peak .

The data given above suggests that a console should be provided for each dispatcher work station in a manual system; some load within the hour).

suggestions and recommendations regarding the desirable fea- e Average and maximum length of time from start tures of such a console will be found in Chapter 6.

of an incident to the time it is closed (by incident priority).

De procedure for making work load measurementsin a

.syctem without CAD is essentially the same as described above, e Number of patrol units assigned to the dispatcher.

= That is, a typical busy period, or several such periods, should be analyzed.This is most easily done by replaying a tape record. ,

ing of the traffic to and from a given dispatcher, and using a A review of the data novided by such an analysis will be topwatch to determine message lengths. Items that should be useful in evaluating the fea:ures offered by different CAD sys.

I aovered in the analysis include: tems and in designing the system itself. Such information, .

especially that related to message type,is helpful in the selec. .

'=

e Number ofincoming messages and length of each. tion of keyboard functions and formats for screen displays.

The data on dispatcher utilization and number of simultaneous e Number of outgoing messages and length of each, incidents being handled are the principal measures of work load, although the number of patrol units assigned to the dispatcher e influences the work load. As noted earlier,a CAD system does Case number associated with each incoming and not necessarily reduce the work load as measured in these terms, outgoing message (where messages are related to specific cases)- but should make a given load less stressful for the dispatcher to handle.

Type of message (status change, assignment of a patrol unit, data- file query, monitoring, and support). 5.3 Systerri Design Decisions e Number of cases initiated during the observation Once a department has begun considering adoption of a period.

CAD system and has carried out the first two steps of defining e functional requirements and analyzing the work load, the next Number of cases closed during the observation step is to make some of the essential decisions about what the petiod.

system willlook like and how it will operate, ne purpose of

~

this section is to review each element of a CAD system and

[ From . such a set of observations, it 'is possible to outline some of the considerations tha: should influence such

determine some of the parameters ofinterest for a CAD system decisions. Wherever possible, specific data on the system

-(as well as for general management purposes): element is presented.

( Length of messages by type (status change, assign.

ment to incident. requests from patrol unit.etc.). 5.3.1 Trunk Line Requirements Itate of messages (incoming plus outgoing). uns A police command and control center, manual or with is done by. dividing the observation period into CAD, will in most cases have a complaint board operator l 20 I......._. ' - "" - -

. s l

1

(

l

)

(C1101, although ceriam departmems combmc ihis poutmn he procedure for descammmg the number of trunk lines with that of dispatcher (Glendale and Palm itcach County in and operators needed is as follows:

our survey) in either case, the function of taking calls for service from the pubhc must be performed. If a new CAD (1) ne planner :pecifies the performance level desired, system is to be implemented, it may be a good idea to in terms of what percentage of calls are allowed to reexamine the CDO position to determine how wellit is per-receive a busy signal, and what mean waiting time forming its function and whetner any modifications should be is acceptable.

made at the time the new equipment is installed.

This section will discuss methods of determining the (2) The planner measures (or estimates) the peak call required number of trunk lines to service the CBO position, and the numbers of primary and secondary eperators required

~.

to maintain a given level of service. (3) Ths design curves shown in this section are used to derive the required numbers of trunk lines and A major system dtsign decision is that of whether or operators.

not to have a secondary operator position. Figure 4 shows the typical now of calls through a complaint board system having Typical performance specifications are in the following a secondary operator position, although the first part of the form:

flow is the same in either case. Calls from the public are auto-matically connected to an available trunk line as long as they (1) No more than 5 calls out of each 1000 shall receive are not all occupied (in this case the caller receives a busy sig- a busy signal. ~

nal). He incoming calls all go to the automatic call distributor system, which attempts to find an available prituary operator.

If none is available, the call is placed in a queue until a primary operator is free to receive the calt Primary Perator queue shall not exceed M sec-onds.

Those systems having secondary operators use them to handle the longer calls, generally those a minute or longer in (3) ne average waiting time for the secondary operator duration. ne primary operator receives all calls first and shall not exceed JO seconds.

decides which should be transferred to the secondary operator (or to some other city department). In all cases the incoming he numbers entered in the " blanks" above are typical; call continues to occupy a trunk line until the caller hangs up, the planner should establish these for his own system on the even when the call is transferred to the secondary operator or basis of his own measurements of peak loads, and his estimates another department. of how frequently a given call rate might be exceeded.

numas, con 50tt fauna LIN() pg,uAgy

- ' ~ ~

CoNiOt t CoM t gg C at t _ _ (att Mia4Aa v Ol\teemnot con \otr MCoNoaav

_ _ _ _ , CON W Pan =Any C ONtOsi n uaa e y C ONtoti Fg4 Complaint board operator c ell flow i

21

[

As part of his measurement of call rate,the planner should message flow to the original 0.5 perecct busy signal specifica.

have collected data on call durations. lie is then in a position tion becomes 14 once more. With both changes (5 percent

[

to determine peak loading on the trunk lines as follows: busy signals,120-second calls), the number of trunk lines Peak call rate (calls / hour) X call Trunk work load " duration (sec) 5.3.2 Primary Complaint Board Operator Positions As an example, suppose that on a busy Friday night a The procedure for determining the number of primary police department is receiving 200 calls per hour and the calls e mplaint board operators needed is similar, but in this case are serviced in 150 seconds on the average, Service time includes the performance parameter is the average waiting time before

/*d !atal time the call was in the system, including any time the caHu as conucted to a primary opuator (r.e., the avuage se, ta queue waiting for either the primary or secondary length of time a call remains in the ACDS queue before the cperator (plus the time spen'. talking with either operator or ACDS can find an available primary operator). This delay will with another city department). In this case: naturally be shorter if there are more operators, but more operator time would be spent waiting for calls.

200 X 150 Trunk work load = = 8.33 load units For our purposes, delay units are obtained by dividing 3600 the average waiting time by the average measured service time (time the operator takes to handle the call, not including wait.

. Now the curves of Figure 5 can be used to determine the ing time):

number of trunk lines needed to bandle this load with the per.

formance already specified.nc value of C (the number on each mean waiting time in seconds curve)is the number of trunklines. Delay = mean operator service time in seconds

i. Referring now to Figure 5, we note that our requirement ne operator work load is then cr.lculated as:

for not more than 5 calls per thousand receiving a busy signal (3in, translates to 0.5 percent, which is the lowest horizontal line, fpeak call rate (calls per hour))

Fsilowing along this line to the value of 8.33, we take the l X mean service time in L nearest value of C to the right,which is 17.This is the number '## "*

Operator work load a of trunk lines needed to meet our specifications under uur 3600 peak load conditions. ,

Returning to our previous example, we have calls arriving it is of intercshto note the effect of changing some of at the rate of 200 per hour and the average service time per call the parameters.'For e . ample,if the percentage of calls receiving is 100 seconds. We have specified that the average waiting time busy signals were increased from 0.5 to 5 percent, the number shallbe no more than 2.5 seconds,so that our number of delay 1 cf trunk lines needed to meet the same demand becomes 13. units is:

And if the average call duration becomes 120 seconds rather than 150 seconds, the number of lines required to handle the Delay unit = 2.5/100 = 0.025

, e . ema or nu.* u *s The operator work load, calculated as shown, becomes:

so hp M%NN73;._ .j- Work load = 200 X 100/3600 = 5.56 load units

~ Yf $C<S$ 7 jp%

- h

3*

f j/g/Ky/ /77" <

Now we turn to Figure 6 to find the number of primary operators required to handle this load with no more than the g/

/ / f

• s 1 ..

/ specified average wait. The 0.025. delay. unit line runs horizon.

2-f~ -

7 j  %(( tally near the bottom of the figure. It intersects the vertical 5.56-load. unit line at the point shown, and we take the nearest curve to the right as before. This is the curve for 10 primary

-)

.i .

7 / /

j- 7--

/

+

( o.s e

i

/

i e s e 3

i e , .,a operators.which is the number needed.

Once again,it is interesting to note the effect of changing Fig. 5. Trunkline *sion the parameters. If the specification for average waiting time is 22

x ~

1

{

)

\

I r

k j changed from 2.5 seconds to 5 seconds, for example, the delay We calculate the number of secondary operators needed I unit value becomer: in the same way as for the primary operators, calculating the delay and the work load and finding their intersection point 5/100 = 0.05 umt on Figure 6.

and the intersection of this value with the same work load falls First, however, we need to define what kinds of calls are between the curves for 8 and 9 operators, rnaking the required to be handled by the secondary operator. We will de;ine "long" nurnber 9. And if the average operator service time drops fro.n calls as those that require more than 60 seconds to service and 100 seconds to 60 seconds, the work load becomes:

which do not require dnpatch. All other calls (namely those that either require dispatch or take less than 60 seconds) are 200 X 60/3600 = 3.33 units defined as "short" calls.

and ordy 7 operators are needed to rnaintam the 2.5 second in our sample specification, we indested a maximum average waiting time.

average waiting time for a secondary operator as 20 seconds.

For our example of a department with calls arriving at the rate 5.3.3 Secondary Complaint Board Operator Positions of 200 per hour, let us assume that 5 percent, or 10 calls per hour, are "long" calls that are to be transferred to the second.

Not all departments use a secondary complaint board ary operator.We will assume further that the mean service time operator to handle the longer calls, but some have found it a of the secondary operator is 5 minutes (300 seconds). Now our good way to improve service to the pubhc and reduce the work delay calculation is:

load on the primary operators. In calculating the number of secondary operators needed, we note first that all calls going to a secondary operator must first go through a pnmary operator.

and <>ur work load calculation is:

If all secondary operators are busy, there is a waiting time q that must be added to the waiting time required for the caller Work load = 10 X 300/3600 = 0.83 unit to reach the primary operator.'the waiting time fo: the second-ary operator will naturally depend on the number of secondary On Figure 6, we find the intersection of these two values in the operators.

area between the curves for 2 and 3 operators, meanmg that we y '

10 II

. I t NUM4(B (# OMtAions 3 4

0.5 -

3 c.4 -

1 E

b 0.3 >-

r b

I 7, ,

ifCOHo atY  !

(to i

-@ . Peimasy

,s -

m_- . ,

o I e 4 4 40 UN 8% of woes loa 0 Fig 6. Cornpleant board operator position deseen 23

. _ _ - - - . . . . . . . - . . - -